Professional Documents
Culture Documents
Fundamentals and
Special Problems-of
Synthetic Aperture Radar (SAR)
LsAspects Fondamentaux et
Thismatria
pulictio
inthi wa asembed o sppot aLectureSeries~
udrtesponsorshipof the Avionics Panel ofAGYARD and the Consultant
and Exchange Programme ofAGARD presented on Sth-6th October 1992
in Bad Neuenahr, Germany, Sth-9th October 1992 in Gebze-Kocaeli
(near Istanbul), Tu rkev and 26th-27th October 1992 in Ottawa, Canada.
. . - '- - - -. . . ..
Fundamentals and
Special Problems of ....-
Synthetic Aperture Radar (SAR)
Les Aspects Fondamentaux et
les Probl~mes Splcifiques aux
Radars AOuverture Synth6tique (SAR)
92-257 i
....... U3U3U.t.....
The Mission of AGARD
According to its Charter, the mission of AGARD is to bring together the leading personalities of the NATO nations in the fields
of science and technology relating to aerospace for the following purposes:
- Recommending effective ways for the member nations to use their research and development capabilities for the
common benefit of the NATO community;
- Providing scientific and technical advice and assistance to the Military Committee in the field of aerospace research and
development (with particular regard to its military application);
- Continuously stimulating advances in the aerospace sciences relevant to strengthening the common defe,ce posture;
- Improving the co-operation among member nations in aerospace research and development;
- Providing assistance to member nations for the purpose of increasing their scientific and technical potential;
- Rendering scientific and technical assistance, as requested, to other NATO bodies and to member nations in connection
with research and development problems in the aerospace field.
The highest authority within AGARD is the National Delegates Board consisting of officially appointed senior representatives
from each member nation. The mission of AGARD is carried out through the Panels which are composed of experts appointed
by the National Delegates, the Consultant and Exchange Programme and the Aerospace Applications Studies Programme. The
results of AGARD work are reported to the member nations and the NATO Authorities through the AGARD series of
publications of which this is one.
Participation in AGARD activities is by invitation only and is normally limited to citizens of the NATO nations.
ISBN 92-835-0683-9
HI
Printed by Specialised Printing Services Limited
40 ChigwellLane, Loughton, Essex IGiO 37Z
Abstract
The Lecture Series will cover the field of airborne and spaceborne SAR with respect to its technical realization in order to
# convey the participants' ideas and know-how on SAR, on its capabilities and on the technology necessary for the successful
construction and application of airborne and spaceborne SAR systems.
The basic principles of SAR will be explained and its peculiarities, ambiguities and special effects will be highlighted especially
in comparison with airborne and spaceborne Radar with Real Aperture (RAR). The influence of speed and altitude variations,
lateral motions on respective compensation possibilities will be presented.
The antenna is a system related SAR component. Therefore, the influence of the antenna parameters on specification and
capabilities of SAR will be considered. Polarization effects and multi-polarization SAR are presently key-points for SAR
development and SAR application as well as questions on absolute SAR calibration. The advantages, necessities and limits of
these topics will be included.
Digital SAR processing is for SAR indispensible. Theories and special algorithms will be given along with basic processor
configurations and different processing techniques on hardware and software bases.
The simulation of SAR-systems as well as SAR-products will also be a topic of the Lecture Series. This includes the simulation of
SAR-techniques and SAR-components as well as the simulation of SAR images. Special SAR-methods like squint-, stretch- and
spotlight-techniques for example will be presented in addition to the inverse SAR-techniques using the motion of targets instead
of the motion of the radar. A presentation of the state of the art giving examples of presently planned and up to now realized
airborne and spaceborne SAR with its application foreseen will conclude the Lecture Series.
The Lecture Series should appeal to technically and technologically oriented engineers concerned with development of SAR
and to scientists, who have to work with SAR for different applications as well as to students of both specialities who have
already attained a high degree of knowledge in techniques and of remote sensing applications.
This Lecture Series, sponsored by the Avionics Panel of AGARD, has been implemented by the Consultant and Exchange
Programme.
iii
Abrege
Ce cycle de conferences traitera du domaine des SAR aeroportes et spatioportes du point de vue de leur rtiaisation technique.
Les conferenciers mettront a profit leur competence POUr presenter leurs Wes concernant les SARl. leurs possibilit~s, et les
technologies qui sont Autiliser pour r~ussir l'industrialisation et la misc en oeuvre des systemes SARl airoportis et spatioportes.
Les principes de base de ces; syst~mes seront exposes, ainsi que leurs particularites, ambiguitis et sp~cificitas, en particulier par
rapport aux radars aeroportes et spatioportes a ouverture reite (RAR). Linfiuence des diplacements lateraux ainsi que des
variations de vitesse et d'altitude sur les possibilitis respectives de compensation sera 6galement presentee.
L'antenne est un composant du systime des SAR. Par consequent, l'infiuence des parametres d'antenne sur les sp~cifications
techniques et les capacites des SAR sera prise en consid~ration. A l'eure actuelie, les effets de polarisation et de
multipolarisation SAR doivent &tre consid~r~s comme des points c16 pour le developpement et les applications des SAR. de
m~me que leur 6talonnage absolu. Les avantages, les nicessitis et lea limites de ces donn~es seront examines.
Le traitement num~rique est indispensable aux SAP,. Des theories et des algorithmes sp~cifiques; seront proposes, ainsi que des
configurations de processeur de base et diff~rentes techniques de traitement sous les aspects matiriel et logiciel.
La simulation des systemes SARl et des produits SARl constitue un autre sujet de ce cycle de conferences. Ce sujet comprend la
simulation des techniques et des composants SAJl, ainsi que la simulation de l'imagerie SARl Des methodes sp~cifiques au SAR,
telles que le deport antenne (squint) lea impulsions itaI~es (stretch) et le mode tlI~scope par exemple, seront pr~sentis en
complement des techniques SAR inverses faisant appel aux mouvement des cibles au lieu du mouvement du radar. Le cycle de
conferences; se terminera par une prisentation de N'tat de l'art dans ce domaine, avec des exemples de syst~mes SARl
aeroportes, spatioportis existants et projetis et des applications pr~visibles pour lesquelles ils ont &6econcus.
Ce cycle de conferences est susceptible d'intlresser les,ingenieurs travaillant sur le d~veloppement des SAR et lea scientifiques
appeles a travailler avec les SAR pour diverses applications ainsi que les etudiants ayant deja des connaissances avancees dans
les deux domarnes des techniques en question ainsi que des applications en ulledtection.
Ce cycle de conferences est pr~sent6 par le Panel AGARD dAvionique; et organis6 dans le cadre du programme des
Consultants et des Echanges.
iv
List of Authors/Speakers
AUTHORS/SPEAKERS
Dr John C. Curlander
Jet Propulsion Laboratory
4800 Oak Grove Drive
Pasadena, CA 91109
United States
Dr Anthony Freeman
Jet Propulsion Laboratory
4800 Oak Grove Drive
Pasadena, CA 91109
United States
Mr Jean-Philippe Hardange
Thomson-CSF
178, Boulevard Gabriel Peni
92242 Malakoff Cedex
France
DLR
Institut ftir H-ochfrequenztcch.-.:k
8031 Oberpfaffenhofen
Germany
4 Dr Herwig OttI
DLR
Institut ftir Hochfrequenztechnik
8031 Oberpfaffenhofen
Germay
Dr R. Keith Raney
RADARSAT Project Office
Canadian Space Agency
110 O'Connor St, Suite 200
Ottawa, Ontario K IA IA I
Canada
V
i if Contents
Page
Abstract i
Abrfti iv
List of Authors/Speakers v
Refrrene
Basic Principles of SAR I
by W. Keydel
SAR Simulation7
by D. Hounani
Multi-Frequency Multi-Polarization Processing for Spaceborne SAR 8
by J.C. Curlander and C.Y. Chang
Bibliography B
- - Vi
*I-I
For remote sensing purposes the resolution of the bandwidth which estimates the shortest proces-
the respective sensor is one of the main fac- sed pulselength. The azimuth resolution is depen-
tors. Resolution in the wide sense is defined dent on the antenna aperture 1 which determines the
as the degree to which a sensor can discrimina- half power beamwidth of the radar beam.
te two closely spaced targets, having similar The basic idea of the synthetic aperture radar
properties (geometry, colour, velocity, etc.)
However, angular resolution is a matter of do- (SAR) is the construction after data collection of
minant concern. It is defined as the minimum a very long antenna along the flight path (here as-
anlar sepaationn.
angular separation twefe two itemicn
between to items which can sumed as strongly linear) by means of aurg
si.Alnthfigtphaete data proces-
be distinguished by a system, sing. Along
points the flight phase
for amplitude, path and
are frequency
the measuring
of the
Note that for all systems using elecromagnetic backscattered signal. In this way the real antenna
Notes, thla oall
s stes using heragnetc acts as individual elements of the large (synthe-
waves, the laws of optics apply. The angular tic) array antenna. The stored echos are combined
resolution of an optical system is principally truhdt rcsig n h A mg spo
limited by both the aperture diameter and the through data processing, and the SAR image is pro-
turbulence of the wave propagation medium (i.e. duced.
the earth's atmosphere). Quantitatively the an-
gular resolution
1 is found by the ofratio
an aperture of a given
of wavelength size
X over setv
The Fig. 1emnlg fSRgeometry and the re-
and 2 show the
this size. ... , ,
z
Bmin12! (6)
; d Oepression Angel This T., is simply the time necessary to shift the
Doppler frequency from the absolute highest value
Nadi p, to the absolute lowest value within the bandwidth
Sof
\\\ the filter.
change Fromtime
of fD with (3) and (4) results for the
I ' AD 2u dx 2 u2
1 "R dt I - R (7)
Rn \AT--
2 2
B 2 U
-
ry Y B > ---
X (8)
- - -Swath Width- - -
With AfD = Bm.n results from (5)
(B)2= 2U )2 (9)
.' /
x. ' j.eI Rano. Bin This is the optimum resolution obtainable by use of
a fixed filter for a certain distance R (slant
Rrange). Remarkable is the angular resolution which
is dependent on the square root of R and not R as
is the case for a normal SLAR. The resolution is
Laet Rang. Bin independent of the antenna dimensions.
Principally, for image construction it is necessary
RealAntenna
Footpant to use for each range bin an other bandwidth (Fig. 1,
Pfozaessd Pixel in Y 2). The use of a filterbank following these equa-
theRangeBin tions within the swath from near range to far range
Mn S R S Rf) allows a relative simple real time
processing.
Following (1) to this resolution corresponds a syn- theoretical reolution of so-called focussed SAR.
Folloianta at re Ld
thetic antenna aperture Lfp: o aRemarkable, that it becomes as better as smaller
the real antenna size 1 is. This is opposite to
2u real aperture radars and most optical systems. The
Lf (5) remarkable fact is that for SAR the theoretical li-
mit of azimuth ground resolution is given by the
half antenna length in the flight direction of thet
In order to optimize this method, the smallest pos- radar.
sible bandwidth B,n has to be estimated. Bm. n de- r
pends on the maximum observation time Tmax availab- The independence from range and wavelength is a
le. further remarkable attribute of SAR resolution. A
look on Fig. 1 and 2 shows: For each distance, each
1-3
range bin, holds an other synthetic aperture and these tude signals must be stored as well as in the unfo-
anetc cussed case. But before the addition of the various
synthetic aperture increases with increasing distancethe phase
as well as the geometric
apruedecreases with increasing of any optical
resolutiondistance, d n a eR0 -R
The range difference made. Fig. 3th
hashas totobe eqalizig phe
shows the
aperture of th incresin isae . The impulse response for a point target for different.
independence of the resolution is a reason for the phase errors (30*, 900, 1500). The signal degrada-
possibility to extend SAR results gained with airborne tion with increasing error evidently can be seen.
systems to spaceborne systems.
As allready mentioned the construction of the syn- The consideration of SAR as a synthesized array leads
to the same equations and results with respect to res-
thetic aperture for a SAR can be considered as the olution, synthetic aperture length etc. as the Doppler
artificial construction of an array by means of consideration does. However, the meaning of focussed
computer techniques. The signals will be stored and unfocussed becomes more understandable in the case
correctly with respect to amplitude, phase and the of the synthetic aperture consideration. Also the name
appropriate positions. During the image processing SAR becomes here more evident.
procedure the stored signals will be added up cor-
rectly and processed to a SAR-image. However, it 3 THE PHASE INFORMATION. KEY TO SAR [1 to 5]
will be a difference between a real array aperture
length and a synthesized aperture of the same 3.1 THE PHASE REFERENCE
length. Whilst for a conventional array the one way
beamwidth Ph estimates the resolution, for a syn- The range R between the platform and a point target
thetic array due to the coherent illumination the can be written:
two way beamwidth has to be taken into account,
since the beamforming phaseshift is introduced on
both the paths to and from the target. Therefore, R2 = R + (u-t)2 (20)
for the azimuth beamwidth of a synthetic array, a
factor 1/2 must be included in the formula (1)for
the one way pattern. u = flight velocity and t = observation time (the
maximum observation time is the dwelltime), R. is the
The time during which a point target is illuminated shortest distance between platform and target. With
by the radar beam is called the dwelltime TD (11). R + Ro - 2Ro follows:
2
An illumination time T D can be provided with an 2 - o U t (22)
oneway antenna beamwidth X R,
= P.
1ha• (15) The two way phase change will be twice:
u2 2
t
There is principally a need for some phase compen- = 2x - . (23)
sation during the whole dwelltime. The distances Ro
of the far end of the synthetic aperture to a point This equation is very important as a reference
target are larger as the distance of the middle Ro. -function for digital SAR processing. The phase
(Fig. 1). A so-called unfocussed SAR ignores these shift is a quadratic function of time. This is
differences. This case limits the observation angle the so-called phase history of a point target,
to the area, where the differences in distance are the phaseshift versus time is a parabola.
smaller than X/16, which is identical with x/4 ra-
dian in phase To the quadratic phase function of time belongs a
linear frequency shift. The Doppler shift fD of the
IR. - Rol S - (16) signal due to the inherent constant platform motion
16 is given by
If this focus condition is fullfilled the occuring 2
degradations and the reduction of the result is not OD = 21 fD d 4u (24)
very large in comparison with the ideal case. Howe- -
ver, the dwelltime is not as large as it could be For a transmitted signal Vt = V, sin(2x ft) the re-
and the resolution is degraded as well as the maxi- ceived signal has the form
mum aperture length. From fig. 1 results with c
P, + Ro - 2Ro:
Vr = V2 sin(2% ft+ 2x fD - t) ,
2
L~)2 = Rj - P. - 2R,,(R - R). 7)Vr=Vsi2xf+2 U t (25)
2 Vr = V 2 sin[2(f + 7-R. -)t]
The limitation (16) leads to the maximum aperture
length of a unfocussed SA: This is strongly equivalent to a linear frequency
modulation. The received signal of a SAR is linear-
(18) ly frequency modulated. This modulation will be
2 .considered as a code which designes all points with
respect to their azimuth angle during flight time.
This aperture is responsible for the resolution The steepness of the so-called chirp is
=ii;__
obtainable and leads to
2
2v
rau, (19) R,
The bandwidth required for this linear modulation
In the focussed case the incoming phase and ampli- results from the dwell time T0 :
__
iI
1-4
Bf = 2 TD (26)
AR AR 2
Fig. 4 Example for focus influence of an image quali-
ty. The lower image is unfocussed (errors re-
sult from velocity) the upper image is focus-
2 (29a) sed.
DOF = 2AR0 = 2X L
4. HIGH RESOLUTION IMAGING BY USING THE PHASE
With (12) results REFERENCE FUNCTION
8 r. The phase of the received signal (25) contents the
DOF (29b) required information on an observed target. These
information can be extracted by a quadratic demodu-
The depth of focus becomes smaller as ra is made lator which eliminates the terms with the carrier
smaller. This is a very important fact for image frequency and leaves the socalled "inphase compo-
processing. For example, if X = 5 cm, ra = 3 m the nent I" and the "quadrature component Q" and a
DOF is about 1.44 km and, therefore, the processing joint amplitude factor A containing amplitude or
of a 4.3 km swath requires 3 different reference signal to noise informations on the target.
functions. For a resolution of 50 cm under the saae
conditions the DOF is 40 m and 10 references are u2 t
required. This increases the processing complexity. I(t) = Acos(2x( -)t)
Fig. 3 and Fig. 4 show the influence of focussing (30)
effects on SAR images exemplarically. u2t (
Q(t) = Asin(2x( )t).
1.6.
:rM../
,AU +T/2
So = J S.(t)'Sr(t+tr) dt , (32)
Fig. 3 Point target impulse response degradation for -T/2
unfocussed prozessing with different phase de-
I
viations (300, 90@, 180").
2 2 +T/2
Soe 2x X e,
2 f e- 2I
2x
2 U2 Rtr t dt,
-T/2
! 1-5
2 u S sinc{ U t) . 3) zes the vector character and the signal delay time
Principally, radar uses information on electroma- Fig. 7 shows the signal to noise ratio3 as a func-
gnetic waves characterized with amplitude E, phase tion of range [7]. The dependence of R and of the
9, frequency f, the polarization which characteri- antenna diagram becomes evident as well as the so-
called amplitude line-echo (ALE) which results from
1-6
40
30
20
10
0 100 200 00 50 W
0
_ I
00 20 0 400 WO
reflections directly underneath the aircraft. This ment of one pixel only. Principally this speckle
can be reduced by image averaging with the use of
curve can be used for image correction. multi looks for instance, this reduces the standard
Fig. 8 gives the signal intensity curve of Fig. af- deviation proportional to the root of the look num-
ber and smoothes the speckle. Speckle can be seen
ter correction. The image in Fig. 6 is corrected ( in all the radar images presented here.
by using the results of Fig. 7.
6.2 SPECKLE STATISTICS The Rayleigh distribution for the envelope Vr be-
longing to (35) is
However, equation (35) shows the statistical nature V2
of the radar response signal due to the target and =
scatter statistics. Normally complexe radar targets P(Vr) expf- -) for 0 f Vr . (40a)
a 2o
like cars, trees and areas etc. have not only one
t 1-7
Var(vr) Vj - V 0.429 .2 V2 %
Reolutfon: 8.94 m
ryred
NY =
R1 + Ax
rrad = (50b)
SAR
RADAR BEAM
,N RADAR-IMAGE PLANE
N ... RADAR-IMAGE
- FORMAT
S!AlT PAJCS----
* RADAR BEAM
a RADAR - MAGE
/ IO
SLOPING SURFACES
SAR
RADAR BEAM
* RADAR-IM~'kGE PLANE
'N Nb RADAR-IMAGE
FOII
/, / C
NNEAR
RANGE ~ .- FARRANGE
,,
SLOPING SURFACES
RADAR-IMAGE PLANE
Table 1 Radar- and contrast equations for real aperture radars and SAR.
2-2
(C/N) = received clutter to noise ratio, (S/N) =of wavelength and distance.
received signal to noise ratio, Pav* = average 2. A better azimuth resolution can be reached with
transmitter power, G = antenna gain, I = wave- smaller real antennas and not with larger an-
length, R = distance, a = radar cross section, 00 tennas as it is the case with RAR and optical
= radar cross section per unit area, k = Boltzmann systems respectively.
constant, T. = 290 K, B = receiver bandwidth, F =
receiver noisequeny,
figure,= fp =t~t=
ulseengh,
quency, Vp = pulselength,
pulse repetition
Lto t = loses
losses, u ~-
fre-
u = SAR- 3. The S/N for a SAR is inversely proportional to
tepafr eoiy
velocity, 1 = real azimuth diameter of SAR anten-
na, ry = ground range resolution, ra = prozessed 4. The S/N for point target detection is inversely
azimuth resolution. proportional to the third power of range and
inversely proportional to the prozessed azimuth
Table 1 shows for point and area targets as well relyopr
ra-
the peculiarities of SAR against conventional
dar. It shows evidently the importance of the pul- 5. The S/C for point targets is independent of di-
se repetition frequency PRF and platform velocity. stance R.
For the point target case the dependence of azi-
muth, resolution, bandwidth, pulselength, wave- 3. AMBIGUITIES
length and distance becomes evident while for area
targets the real aperture length plays an impor- Ambiguities play an important part in all radar consi-
tant part. Especially the expressions in the derations. For all pulse-Doppler radars ambiguities
"Improvement-Factor-Line" and in the "Contrast- exist due to the periodical structures of the signals.
Row" show factors, which are most important for a This illuminates the fact, that the pulse repetition
comparison between RAR and SAR. frequency PRF, fp, will be the decisive factor. Howe-
ver, ambiguities can also be responsible for the choi-
In order to increase the understanding of table 1 se of the basic radar frequency. The principal Dopp-
it should be mentioned, that there are basically ler frequency system received with periodically pulsed
two SAR-techniques called the focussed and the un- Doppler radar for moving configurations is shown in
focussed techniques. Fig. 2 schematically.
Focussed SAR
2
C fp sampling. With bandwidth B = AD fD follows from
Uu 5 (2) (1) the so-called oversampling ratio
c1
R 5 2.
iT . (3) Figure 2 shows the spectral parts coming from the
first and second ambiguous band and contributing to
(This holds for a periodically pulsed radar and for the desired SAR band schematically [5].
small t.)
A combination of (2) and (3) leads to the ambiguous The described ambiguity is mainly PR? conditioned.
product However, a main influence has the antenna pattern
which is partly responsible for the spectrum shape,
C2 1 Fig. 2. Especially improper (that means higher) side-
Uunam unam 5 8 " f (4) lobes enable the SAR to receive power from positive
and negative squint angles, which is within the de-
This equation shows that the choice of frequency for a sired frequency band. These ambiguities are called
pulse doppler radar limits principally the possibility dopler squint angle ambiguities.
for the simultaneous measurement of distance and velo-
city of a radar target. On the other side a frequency Fig. 3, 4, 5 and 6 show the influence of PR? on azi-
limit is fixed if the unambiguous values of Uun. and muth ambiguities. Fig. 3 shows a 2-dimensional re-
R... are given. Example: presentation of a radar image of 5 corner reflectors
and Fig. 4 gives the same image in 3-dimensional
uuna, = 300 m s-', = 40 km, representation taken in X-band with a PRF of 952 kHz.
Fig. 5 and 6 show the same scene taken with a PR? of
frequency requirement: f !51 GHz. 238 kHz. The ambiguities can be seen evidently.
2
u
-." ... ..
. . .;
.....
...
.. ...
.. .. ._...
i - . ._...._ . "_ _... _...
2-4
TP + 2--
azimuth [mJ
Fig. 6 3dimensional SAR image of Fig. 5. The con- -t
tours on top represent the 3 dB values for
the mainlobes and the -20 dB values of the
azimuthal ambiguity sidelobes.
at any time, it can be incompatible with the condition
for the azimuth ambiguity and in this case one makes
allowance for range ambiguities and tries to suppress
these ambiguous signals with proper antenna pattern
design or with special processing procedures respecti-
vely. Fig. 7a shows exemplarically an X-band SAR image
with ambiguities. The ambiguities are eliminated in
Fig. 7b. Fig. 7b SAR image of Fig. 7a, correctly prozessed to
4. PRF-CONSIDERATIONS the squint angle 6.80; all ambiguities are
eliminated (measurement Horn, Moreira).
4 ~~~sive
It hasfactor
been already mentioned, that
for SAR-ambiguities. the PRFfrom
However, is a deci- viewing
other geometry is fixed by other reasons (given or-
viwngemtysfxdbyohreans(vno-
sienftofor SAN-aiguis.deep owever omus otr bit height of a satellite, fixed incidence angle etc.)
point of view the
the
ofaPRF hasaso.
ffetivs
SR deep going consequences
he efiitio ofPRFfor tePFms etndt
the PRF must be eetdvle.Fg
tuned to selected hw
values. Fig. 8 shows
the effectives of a SAN also. The definition of PRF allowed PRF-bands for a satelliteborne system with a
becomes difficult due
andwich to different other conditions
aveto b fulfiled.fixed depression angle, taking into account the varia-
imiatios
tion of the local orbit height [7]. The white ranges
are the allowed PRF-ranges, m is the number of the
There are principal limitations due to azimuth ambi- respective range ambiguities. The lower limit for the
guities, range ambiguities, swathwidth, complete co- PRF with respect to azimuth ambiguity is also given in
verage. There are also unallowed PRP-bands due to geo- Fig. 8.
metrical variations like earth curvature, orbit ex-
centricity, height variations, and altitude line The result is:
echos.
The ambiguity equations (6) and (8) lead to the follo- - with increasing range ambiguity decreases the re-
wing m iiteations nspective width of the allowed PRF-bands.
This can lead to the request of switchable PRF for
2u 1 different purposes (shown in Fig. 9 as necessary due
S Run "to ALE-influences). Substituting (9) into (6) (consi-
+ - dering the equality in both relations) one obtaines
together with the maximum azimuth resolution relation
The choice of PRF estimates the maximum swath width R9 an important relation between swathwidth, azimuth re-
or vice versa. Principally, an impulse needs for cros- solution and SAR velocity:
sing the swath the time i = 2R.1c. Herefrom results:
2u P,
(9 -r, = c . (10)
f1 2R9
..
--
9
A pulsed radar using a single antenna is normally un-
able to receive during the duration of the transmissi-
on pulse, it is blind at certain slant ranges. If the
2-5
i
r ~platform.
PR
z
PRF-Bands Satellite X-Sandpltom
S For a pulsed radar ALE considerations will be identi-
s0o cal with altimeter-considerations. An example for PRF-
4limitation due to ALE shows Fig. 9 in comparison with
-Fig. 8. For m = 12, the influence of ALE leads to the
request of PRF-switching as mentioned in the previous
2500 section.
5. ANTENNA PECULIARITIES
rn.l6 Equation (6) gives a relation between the real azimuth
2000 diameter of the SAR-antenna, flight-velocity and the
1 20 30 40 50 60 70 80 90 RP R - X
Ah/km - = inODi - sini lyert (Ii)
Fig. 8 Allowed PRF-bands versus altitude variations From (6) and (10) follows in connection with (1.1):
for a satellite, X-band.
c sinb i 5 lvert laz " (12)
SATELLITE
X-Band Equation (12) gives the lower bound for the antenna
area.
SPRF 6. PECULIARITIES AND CONSTRAINTS CAUSED BY PHASE ERRORS
m-14 200 Z Essential for the effectivity of a SAR system is its
210 phase coherency. High quality images can be produced
only if the targets phase history is observed along a
M-13[ precisely known radar translation. However, these pha-
o se history is normally not known exactly. Phase errors
917M usually occur. Principally there are two different ty-
m-12 pes of phase errors:
1 More or less deterministic errors, caused by well
m-11 known geometry effects and defined instrumental in-
fluences.
SAmftgmity-nift 1Ig4H In lesson 1, Cap. 3.2 the "Depth of Focus" has been
l MW =considered and in Fig. 3 an example for the influence
XnI ,' of defocussing on a SAR image has been shown there.
-20 -10 10 20 The same considerations can be used for the computa-
LAh/km -- w tion of the so-called "Range Curvature Effect". The
range curvature (RC) is given by
Fig. 9 PRF limitations due to altitude variations
Ah, example for PRF switching. RC = - . (13)
*The Altitude Line Echo
For high azimuth resolution application a long in-
The allready mentioned altitude line echo (ALE) is the tegration time is required and image degradation may
radar signal coming from Nadir. The time duration of be caused if the time delay variation corresponding
ALE is with sufficient accuracy identical with the to (13) reaches the same order or exceeds the range
2
transmitter pulse duration %p. ALE appears earlier resolution cell (Cp/ ) for a pulse radar. From (13)
than the desired swath echo. On the one side it will follows
more attenuated than the desired signal proportio- )2
nal to the sidelobe level of the antenna pointing in Cu TD)2
(
Nadir direction and on the other side it will be less RC - 2 (14)
attenuated due to the shorter distance, where ALE is
resulting from. o has normally a higher value for Na- 12 R
dir direction than in any other directions and this RC - - . (15)
will increare the ALE-signal level. Principally, the 16 r.
ALE can be used as a reference for geometrical cali-
bration as well as for altitude estimation of the
2-6
By this the importance of ra is evident. In compari- Other so-called deterministic phase changes can occur
son to ry follows: due to orbit excentricity of satellites and each rota-
X2 Rtion effect, which can normally be estimated exactly
R (16) or due to antenna influences like diagramme deforma-
ry 16 ry r2 tion or angle switching. Both effects can be measured
and estimated exactly.
This equation gives the number of resolution cells
through which a point migrates during the formation 6.2 PLATFORM INSTABILITY EFFECTS
of the synthetic aperture. Ratio values in excess
of 0.3 will normally cause image degradation and Fig. 11 shows measured displacements in the line of
this must either be avoided by design or compensa- sight of an airborne SAR, extracted by a RDM motion
ted during signal processing (2]. compensation equipment.
In Fig. 10a the range curvature clearly can be seen.
However, here the curvature is less enough to avoid
image degradation.
processed image
"
Sideal impuls
-20
. raw da 'a _40 ! d
SAR
raw data -40
A 0
4 01
-0.5 0-. 0.5 15
Tim [,]
Fiq. 10a The raw data received from a point target
(lower image) show as well as the range com- Fig. 12 C-band impulse response of a point target
pressed data the influence of range curva- for DLR-E-SAR degraded by phase errors due
ture. to- displacements and turbulence (RMS = 1 m
s ' , peak value 2 m s - 1, displacements be-
tween -2 m upto 4.5 m).
AIn principle these effects can be neutralized
by
motion compensation by measuring the dynamic beha-
---viour of the platform and correcting the SAR data
0() eighter on board of the aircraft or on ground. Two
methods are principally possible, eighter the mea-
a 0surement using an Inertial Measurement Unit (IMU)
ore a procedure, which extracts the true forward
velocity and line of sight changes and accelera-
tion out of raw data called 'Reflectivity Displa-
cement Method" RDM. RDM, recently has been deve-
S/ loped at DLR [9, 10). This procedure has been used
to exclude the errors caused by aircraft motion and
velocity instabilities during the processing of the
images shown here. In the DLR approach only the in-
formation in the radar signal has been used for mo-
Fig. 10b Impulse response of a point target disturbed tion compensation; principally no intertial platform
due to range migration caused by range curva- or other equipment for measuring details of the
ture. The sidelobes are caused by ambiguity. aircraft motion is necessary. This method enables
The 2dimensional representation on top shows the estimation of deviations of the aircraft from
evidently a smearing effect (DLR-E-SAR, Horn, an ideal path as shown in Fig. 11.
_Moreira)
2-7
uTD )2
R : R2 + (_-
0 2
Remarkable is: The tolerable au becomes more critical for a moving target at a distance R the displacement
if the azimuth resolution is improved. Ax
Usually a SAR signal processor locates the position of ig. 15 SAR image of a sea surface with 2 ships taken
a non-moving target in the image place where its dopp- by SEASAT. The displacement of the ships
ler frequency is zero. If the target is moving with a against the satellite clearly can be seen. The
radial velocity component Utr, then it imposes a Dopp- opposite displacement against the wakes cor-
ler shift on the signal. Thus a target moving towards respondings to opposite velocity directions
the radar will cause a shift Ax in target location in (processing: DLR-T-DA). Mean displacement is
the flight direction of the radar (and vice versa).
if the radar is moving with a velocity U than results In addition to radial velocity, radial acceleration
will also cause image distortions such as defocus in
2-8
Error Source Image Image Azimuth Range Range Azimuth Image Image
Misregi- shift Defocus Defocus Walk Mainloge Sidelobe
stration Loss Increase
Range r !
(altitude t
and cross r
track) dr
Azimuth x
(along ••
track) x
dtx
Propagation 1.1
path j 1
If1
'
Table 3 Effects of phase errors,caused by platform or target motions respectively (the relative mo-
tion is from relevance only), as well as caused by phase jitter A9 in the propagation path
and in the SAR electronic.
2-9
0 0,
-40 -20
-5 - 5
-60 -60
-70i -70
-0.6 -0.4 -0.2 0 0.2 0.4 0.6 -0.6 -0.4 -0.2 0 0.2 0.4 0.6
Tim(] Tim.[.]
Fig. 16a Impulse response, shifted by a linear Fig. 16c Impulse response, degraded by a cubic
phase error, phase error.
0 0
-10 -10-
-20 -20 -
S-30 !.-30
40I -40
-50 -30
-60 -60
-70 -70
-0.6 -0.4 -0.2 0 0.2 0.4 0.6 -0.6 -0.4 -0.2 0 0.2 0.4 0.6
Fig. 16b Impulse response, degraded by a quadratic Fig. 16d Impulse response, degraded by a random high
phase error. frequency phase errors.
-
Principally, the power required increases with the capacity. A reasonable goal would be 250 Mbit s 1.
3rd power of the radar distance. (This means a doub- New data transmission systems with splitted data
ling of the distance, i.e. the orbit height, requires links to data relay satellites or ground stations
an 8-fold multiplication of the required power). A respectively are under preparation.
larger antenna beams more power to a required area
(expressed through the so-called antenna gain) than Data storage capability has to be increased also.
a small one, and, therefore, a large antenna would As a first step new recorders with capabilities
seem to be favourable. But the ground resolution of exceeding 100 Mbit s- 1 are qualified for use in
a SAR is improved for smaller real antennas. These space. The present approach is to use more than
considerations lead to a tradeoff, which must be ma- one recorder, i.e. one recorder for each channel
de carefully in order to fulfill a satellite SAR's in multipolarization and multifrequency SAR, as
requirements. In any case the transmitter power is a on the SIR-C mission.
limiting element for the design of a SAR as well as
the antenna. Todays state of the art are a few hun- Image data processing capabilities also are li-
dreds watt (mean power). This implies requirements mited. However, this processing time will de-
for the primary power supply of a satellite. State crease rapidly during the next few years and at
of the art is about 6 kW to 10 kW. These require- the end of the decade real time processing with
ments can be fulfilled with solar power generators excellent image quality will be possible.
and atomic generators as well. The antenna dimen-
sions for spaceborne SAR have at present values of Requirements for high resolution and wide swath
about 2 m x 15 m. in continous strip map mode are in conflict with
data handling capabilities, and with require-
8. FUTURE SAR TECHNIQUES ments for a large antenna to conserve power.
Therefore, new SAR techniques have to be intro-
SAR sytems produce a tremendous amount of data duced and developed which allow electronically
(examples for data-recorder bitrates on the ground: steered beams. For this purpose the spotlight
ERS-1 n1 02 MBPS, X-SAR/SIR-C G 45 MBPS per chan- mode and Scan-SAR modes are under consideration.
nel M,)DLR airborne E-SARZ 28 MBPS) Requirements The length of a synthetic antenna corresponds
2 ") to the section of the flight path from which
for simultaneous high resolution and large swath one target stays within the antenna beam. This
widths make the data rates higher. All requirements fact leads to the requirement for wide beams
for extensions of SAR to multifrequency and multipola- and, therefore, small antennas for conventional
rization capability respectively entail a multiplica- high resolution systems. The same effect, how-
tion of the data rates and this would exceed the pre- ever, can be reached if a small antenna beam
sent limitations of data handling. This seems to be a (from a large antenna) can be continuously
key problem in all high resolution imaging systems. pointed at the target. This allows also a lon-
Therefore, different requirements have to be fullfil- ger synthetic array and, therefore, a finer
led in order to handle or reduce the data stream of azimuth resolution [11]. However, the gain of
future systems either by means of onboard processing azimuth resolution entails a loss of coverage
or with development of advanced SAR systems like spot- due to the fact that during the continuous
light SAR. One solution is to increase the data links spotlight illumination of one small surface
T2-10
area, the sensor passes other parts of the German Patent P 41 24 062.6.
swath which are not illuminated. Therefore,
the spotlight mode can be used for the enlar- [11] Buckreu8, S. Motion Errors in an Airborne
gement of a sector of the observed swath similar Synthetic Aperture Radar Sy-
to the zooming with an optical camera. The Scan stem.
SAR Mode can be used for an extension of the ETT Vol. 2, No. 6, Nov/Dec.
swath in radial direction using more than one 1991, pp. 655-664.
beam generated in a time shared manner.
method increases the swath at the expenseThis
of [12] Keydel, W. Verification Using Spaceborne
azimuth resolution [16 to 20]. A spotlight SAR Microwave Imaging.
in orbit, however, would be a rather expensive IEEE Technology and Society Ma-
enterprise. gazine, Dec. 1990/Jan. 1991,
The combination of all modes allows variable pp. 53-61.
resolution and swath widths as well; the trade- [13] Horn, R. C-Band SAR Results Obtained by
off between resolution, swath width, power etc. an Experimental Airborne SAR
leads to optimised configurations. Sensor.
9. REFERENCES Proc. IGARSS 89, IEEE, pp.
2213-2216.
[1] Skolnik, M.I. Radar Handbook. [14] Moreira, A. Improved Multilook Techniques
McGraw Hill Book Comp., New Applied to SAR and ZCANSAR
York, 1970. Imagery.
[2] Hovanessian, S.A. IEEE Trans. on Geosc. and Re-
Introduction to Synthetic mote Sensing, Vol. 29, No. 4,
Array and Imaging Radars. July 1991.
Artech House, Inc., 1980.
[15] Moreira, A. A New Subaperture Approach for
[3] Tomiyasu, K. Tutorial Review of Synthetic Real-Time SAR Processing.
Aperture Radar (SAR) with Ap- ETT, Vol. 2, No. 6, Nov.-Dec.
plications to Imaging of the 1991.
Oceans Surface.
Proc. of the IEEE Vol. 66, No. [16] Brunner, A. Concept for a Spaceborne Syn-
5, May 1978. Langer, E. thetic Aperture Radar (SAR)
[4] Kovaly, J.J. Ottl, H. Sensor Based on Active Phased
Synthetic Aperture Radar. The Zeller, K. Array Technology.
Artech Radar Library. AGARD Conf. Proc. No. 459 on
Artech House, Inc., 1976 High Resolution Air- and Spa-
[5] Ulaby, F.T. Microwave Remote Sensing Vol. ceborne Radar, Papers presen-
ted at the Avionics Panel Sym-
Moore, R.K. II, Radar Remote Sensing and posium held in The Hague, The
Fung, A.K. Surface Scattering and Emis- Netherlands, 8-12 Aug. 1989,
sion Theory. pp. 23 Al - 23 A10.
Addison Wesley Publ. Comp.,
1982. [17] Jatsch, W. Concept of an X-Band Synthetic
[6] Elachi, Ch. Spaceborne Synthetic Aperture Langer, E. Aperture Radar for Earth Obser-
Ottl, H. ving Satellites.
Bicknell, T. Imaging Radars: Applications, Journ. of Electromagnetic Waves
Jordan, R.L. Techniques and Technology. and Applications, May 1990.
Wu, Ch. Proc. IEEE Vol. 70, No. 10,
Oct. 1982, pp. 1174-1209. [18] Luscombe, A. Taking a Broader View: Radarsat
[7] Schlude, F. Imaging Radar Systems. adds Scansar to its Operations.
Proc. IGARSS '88 Symp., Eding
Proc. of an ESA-EARSeL Workshop burgh, Scotland, 13-16 Sept.
held at Alpbach, Austria, 16-20 1988, pp. 1027-1032.
March 1981 on "Coherent and In-
coherent Radar Scattering from [19] Raney, K. Canada's RADARSAT.
Rough Surfaces and Vegetated Remote Sensing Yearbook 1990,
Areas", ESA-SP-166. (personal communication).
[8] Moreira, J. Estimating the Residual Error [20] Luscombe, A.P. The Radarsat Synthetic Aperture
of the Reflectivity Displace- Radar: A Flexible Imaging Sy-
ment Method for Aircraft Mo- stem.
tion Error Extraction from SAR Proc. 11. Canadian Symp. on Re-
Raw Data. mote Sensing, Waterloo,
ICCC Intern. Radar Conf., Ar- 22.-25.6.1987.
lington, USA, 1990, pp. 70-75.
[9] Moreira J. A New Method of Aircraft Motion
Error Extraction from Radar Raw
Data for Real Time SAR Motion
Compensation.
12th Canadian Symposium on Re-
mote Sensing, IGARSS '89, Van-
couver, Canada, proc. IGARSS,
IEEE, 89, pp. 2217-2220.
[10] Keydel, . Verfahren zur Extraktion von
Moreira, J. durch die Atmosphire verursach-
ten Phasenfehlern des R~ck-
streusignals eines Abbildungs-
radarsystems aus Radarrohdaten.
3-1
In contrast to this well known effect, the focu- Hence, by the path of the sensor is understood the
sing of a SAR sensor is in principle insensitive path of the phase centre of the antenna. If the
to angular deviations. This is because the sensor path is not measured directly at the phase centre,
resolves the scene in terms of displacements ra- roll, pitch and yaw angles need to be known to
ther than angles. correct for the offset. Errors in these angles
will then lead to degradation in image quality,
In the plane orthogonal to the flight path, i.e. further contradicting the generalisation above
in the range direction, the position of a target that the focusing of the SAR is insensitive to an-
is determined by measuring the delay of the tar- gular variations. For the sake of simplicity, the-
get's echo (range delay). The spatial resolution se angular effects have been neglected below. If
is determined by the length and bandwidth of the needed their influence can be derived from the
transmitted pulse. If the sensor is rotated in formulas provided.
this plane (roll angle), the influence of the an-
tenna gain pattern may cause the strength of the Clearly, a stable flight path greatly eases the
echo, i.e. the image intensity, to vary but the task of motion compensation. Satellite platforms
delay and, hence the spatial resolution will re- fall into this category even though in some cases,
main unaffected. e.g. space shuttle the orbital parameters as well
as the attitude can be poorly defined. The real
Parallel to the flight path, i.e. in the azimuth challenge for motion compensation is presented
direction, the scene is resolved by matched filte- when SAR sensors are flown on light aircraft due
ring the Doppler spectrum, there being a fixed to the influence of air turbulence and the often
correspondence between Doppler frequency and the frantic effort of the pilot to combat it. The air-
relative position of the target on the ground. The borne SAR sensor of the DLR (ESAR) is such a case
spatial resolution is determined by the spacing of [1] and will be used as an example in the follo-
the lines of constant Doppler and the bandwidth of wing analysis. Fig. 1 shows an image of the DLR
the matched filter. Rotation in azimuth produced centre in Oberpfaffenhofen using this sensor which
by pitch and yaw will have no influence on the li- exhibits typical degradations due to motion er-
nes of constant Doppler nor on the matched filter rors. The aircraft was deliberately flown to pro-
characteristics. In principle, therefore, the duce motion errors which can be seen as blurring
of the image (top right) and geometric distortion
3-2
%
D
P$
Alternatively,
OA(t) = L (5)
L
• where Ro is the range at the point of closest ap-
proach of the sensor of the target.
window
If the weighting function is a rectangular
W(t)= WR(t), with
WR(t) = 1 for Itl ! T/2 and
(16)
A' A WR(t) = 0 for Itl > T/2
The phase error 0(t) is, therefore: iy(t) = 4-s cosD •y(t) (
' (19)
0(t) = - ax(t)'@A(t) - Ay(t) 'cose + -X
+ Az(t)sinD]. (10) - sin " (20)
From equation (10) it can be seen that the in- The first term in the brackets of Eq.(17) repre-
fluence of the motion errors in the x direction is sents the frequency error resulting from the di-
independent of the depression angle OD and, there- splacement in flight direction. The second term
fore, the the
in theofother
the position
swath, ofWhereas, two directions (y the
target within and shows aerror
quency relationship
linearand velocity component, it beingfre-
between angular pro-
z), OD needs to be known. Also, it is evident that portional V-t/Ro, i.e. the azimuth angle @A. This
the variations of the displacements with respect angle is the sum of the so-called squint angle,
the i.e the angular offset of the boresight of the an-
to time in all three directions is reproduced in tenna, and half the azimuth antenna beamwidth.
the phase error, a linear variation in the displa- Eq.(19) and Eq. (20) also show a linear relation-
cement
and so manifesting
on. itself as a linear phase error ship b
ship n frequency
between Equency error
eror and v elt compo-
an velocity on-
and s on.nent.
Assuming that the velocities are constant, we can
The normalized reference function H(t) from Eq.(1)
is given by: now calculate the phase error 0. across the trans-
2
ikt mit pulse length %P;
H(t) = W(t)-T f'.e2 , -T2 t +T/1 , (11)
*px = *x'cp , *py, = Sy'p , p2 = z'¢p (21)
and W(t) is a weigh-
where T is the aperture time
ting function to suppress sidelobes. Often a Ham- Let us now calculate some practical values for the
ming weighting function (4] is used: velocity components and the displacement in flight
direction, which would each result in a linear
WM(t) = a + (l-a)-cosIt a = 0.54 (12) phase error of x/2, a value which would need to be
LT.achieved to influence the impulse response. The
results are given in Table 2. We will use the pa-
Using equations (1), (2), and (11), the impulse rameters for a typical satellite sensor (ERS-l)
response of a point target becomes: and airborne sensor (E-SAR) given in Table 1.
Ax -22830 -75467 m +1
+ 1 a(0)t
a(0)t+3
+ ..3 (24)
x -9339 -6434 m/s
where p(0) is the position error, v(0)
the veloci-
130 1274 m/s ty error, a(O) is the acceleration error and A(0)
is the derivative of the acceleration error at the
beginning of the synthetic aperture.
z -346 -1274 m/s
2IMPULSE RESPONSE
IZ(t) - A, 27 (22)I
20 15 10 -s5 10 15 20
2 ) - jkt Azmuthm I
x W(,) - ejobcos( xfo .e rd .
-5 -\and
2
-10- - - G (f) =sin eO .%}) GDZ(f) , (30)
-15~ - (M
-20 where GDx(f), Gm (f) and G02 (f) denote the displa-
-20 cement PSDs of the platform in x, y, z direction.
-30 - ,, The expected image quality can be predicted, if
The motion error of a platform is regarded as a Considering all frequencies contributing to the
stochastic process, which is assumed to be statio- linear phase error, which are defined by G (f),
nary and ergodic. Thus it can be described most the variance of the shift of the impulse response
2
completely by a power spectral density (PSD) [3]. OL can be evaluated with
2 1/T
The following Eq.(25) denotes the relation between 2= .R) f
the phase error PSD G (f) and the PSD of the dis- OL2 = 2" • f2 G0 (f)df . (35)
placement of the platform from the nominal track 0
GD(f). Within Eq. (26) and Eq. (27) the factor 4.2 Quadratic Phase Errors
(1/2f)2 corresponds to an integration in the time
domain, which enables the conversion from the ac- Concerning the sinussoidal phase error denoted in
' celeration PSD GA(f) to the velocity PSD Gv(f) and Eq. (31), for yo=0, the quadratic term of the
to the displacement PSD G(f). power series becomes
2
f,,t)
GD(f) , (4-X)2.
•
G
=
(f) (25) 0 0 (t) = 0 2x (36)
The components of the platform motion in x, y and Considering all frequencies contributing to this
z direction are contributing to the total phase phase error, which are described by G (f), the va-
error PSD G.(f). riance o02 of the quadratic phase error, measured
at the end of the synthetic aperture results in
Using Eqs. (10) and (25), we get for the x-axis the expression
The relation between the mainlobe broadening and where p, is the mean value of the resolution of
the quadratic phase error at the end of the syn- one look. The looks were evaluated with a Hamming
thetic aperture Q9 (T/2) was determined empirical- weighted reference function and finally root sum
ly, by correlating a simulated backscatter signal squared.
of a point target with a Hamming weighted referen-
ce function (a = 0.54) and is described by 4.5 Determination of the Acceptable Notion Error
Eq. (39).
The acceptable residual motion error, which re-
(39) mains uncorrected, represents the required perfor-
peff 1 + 0.55 2, t *Q(T/2) 5 x mance of the motion compensation system. The fol-
Pideal lowing analysis is based on the E-SAR system para-
meters given in Table 3 and the image quality spe-
The ideal resolution is denoted as pid..a and the cifications in Table 4.
resolution, which is degraded by a phase error is
0elf. For example, a quadratic phase error of x/2
at the end of the synthetic aperture results in a * platform: Dornier Do 228
mainlobe broadening of 6.6%.
e altitude: 1000m ... 3000m
4.3 High Frequency Phase Errors * forward velocity (nominal): V = 70m/s
It has already been pointed out, that phase errors * maximum slant range: Ro= '000m
with periods below the aperture time T will have e depression angle: %= 0- ... 70°
an impact on the integrated sidelobe ratio or the
image contrast. According to Haslam [4], the re- a antenna beamwidth: 6B= 10.
quired expression is * wavelength: XL= 0.2308m
1
r c= 0.0566m
ISLR - G,(f)df (40) Xx= 0.0313m
1/T
* aperture time: TL= l5.0S
Tc= 3.68s
4.4 The Effect of Phase Errors on Multilook SAR
Images
Tx= 2.03s
The multilook technique was developed to reduce * azimuth resolution (l look): p = 0.66m
the speckle of SAR images. This can be achieved by
dividing the synthetic aperture into overlapping
subapertures, which are finally summed up incohe- Table 3 System parameters of the E-SAR.
rently. The impact of motion induced phase errors
on each look can be calculated with Eqs. (35),
(38), (39), (40), where the integration time of
one look TN has to be taken into consideration. a Integrated Sidelobe Ratio: ISL = -20dB
* A linear phase error has the same effect on a & loss of geometric resolution: < 10%
multilook image as pointed out in 4.2.1: Each * pixel shift: < 50% of one re-
look is shifted in the same direction by the sa- solution cell
me distance.
Table 4 Required image quality.
" A quadratic phase error causes a mainlobe
broadening and a displacement of the looks from 4.5.1
eachoth.Spcfication of the P for an Acceptable
Motion Error
The distance dN between the first and last look A power spectral density of the displacement of
[5] is given by the aircraft can be specified, which expresses the
allowable, residual motion error. The specified
d,_ a-rTNV (41) PSD of the displacement is denoted as GD.(f), the
ar 'measured PSD of the Dornier DO 228 aircraft is
denoted as GD(f). The PSD GD.(f) is split up in
where ar is the nominal Doppler rate, subdivisions, which are characterized by exponen-
tial functions, depicted as straight lines, using
2.V2 a double logarithmic scale (Fig. 5). For computa-
-(42) tion, Eqs. (25), (35), (38), (39), (40) were used.
The coefficients and exponents of these subfuncti-
and Ac. is the deviation from the Doppler rate due ons were adjusted numerically to achieve the re-
to the quadratic phase error *0,which is given by quirements in Tab. 4 for L/C/X-band. GD,(f) is li-
=
mited to a frequency range from 0.001 Hz, which
- O • (43) corresponds to a maximum duration of one pass of
TN x'TN2per
2 1000 boundary
seconds of
or 2.2 16 minutes,
Hz. Frequencies
appoximately up-
above to2.2an Hz
From Eqs. (41), (42), (43) and after replacing the would cause a maximum ISLR of approximately -30 dB
look displacement dp by its standard deviation oN, (X-Band) and are not considered to degrade the
we get: image quality markedly. The result is given by
Eqs. (46)-00) and is also shown in Fig. 5. The
-,00 .. . . ... .
L ,l
.. i.
Tc Tx
1 t... i ....
1-3 0.01 0.1 to
E 5
-2 0 9 3
GD,(f) - 3.41"10-6 Mz .(Tc'f) . , (47)
I/TL S f 5 I/T, -4 I I I_r _
-S -6 -4 -2 0 2 4 a
- 7 "4 5 6
GD,(f) - 4.11"10-8 z "(Tx'f) , (48) TIme [a]
1/Tc !5 f < I/Tx
2
G, (f) - 4.11"10-8 M i/Tx i f 5 2.2 Hz (49)
Fig. 6 Simulated motion
error, derived from the
GD1(f) = Gom(f) , 2.2Hz 5 f < - . (50) measured PSD of the aircraft.
I1p
ded by phase errors due to Fig. 5, compa-
4.5.2 Simulations red to an undistorted impulse response.
In addition to theoretical examinations, the ob-
tained results were verified by simulations, based
on the Monte Carlo method. Therefore, motion er-
(3-8
! 0The
-70 quadratic phase error Oo(t) is denoted as
-0.6 -0.4 -0.2 0 0.2 0.4 0.6
Time [Q(t) - I ab't 2 (53)
. 40 Vc(t) : - 9
-gd (55)
20 c(T/2) = T . (56)
Using a digital phase shifter, the amplitude of flight-path can be performed online in the sensor
the phase error 0(t) can be adjusted only in dis- or off-line in the SAR processor. Thereafter the
crete steps, due to the number of available bits. data can be processed as if the platform moves in
2
The variance a. , respectively the mean power of a straight line at constant altitude and parallel
the quantisation noise is to the swath.
0 20 0 so 80
Thus the ISLR, caused by sampling and holding the
detected motion, becomes
Once the flight path has been determined, it has stortion (runway and taxiways) has been corrected.
to be taken into account during the processing of
the SAR data to ensure an undistorted, well focu-
sed image. The correction for the errors in the
IT-
3-10
(60)
* (t) = - a(0)
Differentiating we obtain:
- =-a(0)
a(t) t . (61)
'0. - --- A-a* * - Hence, the coefficient of the quadratic phase er-
ror a(0) can be del-ermined from the misregistra-
tion of the looks. Note that Ro , V and at are all
I '" ' known.
broad enough not to influence the ground reflecti- 0 100 200 300 400 500
vity function or, if it can be corrected for, the
frequency shift can be determined by correlating SHIrTCz
fREOUENCY
the two adjacent azimuth spectra. If At is much
smaller than the azimuth illumination time, the
frequency shift AfDoppler can be determined very
accurately.
Fig. 15 Correlation of the azimuth power spectra
Fig. 4 shows two such spectra for the E-SAR air- form Fig. 14.
crne sensor operating in L-band with a time offs-
et At of 1.075 s.
HZ
Fgtud VOW0 0. In LOS-drM.
Acc.ler.
{dli F- ou4Vocy
200 " 02. lA10.Il0r0200 In
40
s(fk-1) t I= o -to
o -20
-300 -200 -100 0 too 200 300 D10 NE-OF E-SA* SOSTEM .0TI,-lI COIPE l50TIO. PPC 0S$on
FREQUEN~CY
(HZ)
As an alternative to using inertial platforms, it Proc. IEEE Conf. Toronto, 1983, Vol.1, pp.
would be attractive to use the satellite naviga- 126-131.
tion system GPS, being a much cheaper solution. [5] F.K. Li, D.N. Held, J. Curlander, and
A GPS receiver is used in the E-SAR airborne SAR C. Wu:
to provide absolute position of the aircraft. Ho- Doppler Parameter Estimation for Spacebor-
wever, due to the deliberate errors introduced ne Synthetic Aperture Radars.
into the GPS data by the GPS operator (selective IEEE Trans. Geosci. Remote Sensing, Vol.
availability), the data are not accurate enough GE-23, 1985, pp. 47-56.
for motion compensation of the SAR data and they
are only used for annotation purposes. Note, that [61 J.L. Farrell:
the selective availability degradation particular- Strapdown INS Requirements Imposed by SAR.
ly influences the velocity measurements. The use the USA.
Conf. Dayton,of OH,
Proceedings IEEE 21-25
1984 May 1984.
National
of an additional stationary GPS receiver to com-
pensate for these errors (differential GPS) could Aerospace and Electronics Conference.
provide an attractive method for measuring the Naecon 1984 (IEEE Cat. No. 84CH2029-7).
flight path and is worth investigation. [7] D. Blacknell, I.A. Ward and A. Freeman:
An additional method of improving image quality motion Compensation and Geometric Distor-
was suggested by Chan,the so-called tuned auto tion in iborn S ery.
compensator [10]. This method analyses strong Progress in Imaging Sensors, ISPRS Symp.,
point targets to identify paired echoes. From the- Stuttgart, 1-5 Sept. 1986.
se the high-frequency phase errors can be derived. [8] J.A. Moreira:
The method requires much trial and error and can, A New Method of Aircraft Motion Error Ex-
be regarded as an augmentation of
therefore, only cte--cnae.traction tActo f r rRaw Dat f or E
from Radar Raw Data for Real Time
SAR Motion compensation.
'Table summarises the capabilities of the various Proc., IGARSS Symposium, Vancouver, Canada,
methods. 1989.
[9] J. Moreira:
Method Forward Low-frequ. High-frequ. Motion Compensation SAR-Processing Facili-
velocity errors in errors in ty at DLR.
LOS LOS EARSeL '90, Toulouse, France, 5-8 June
1990.
inertial
measure- yes yes yes [10] Y. Chan:
ment A Tuned Auto-compensator for Residual An-
tenna Motion in Synthetic Aperture Radar
Look Systems.
misregi- either or no IEEE Trans. Geosci. Remote Sensing, Vol.
stration GE-24, Nov. 1986, pp. 1025-1027.
H.bttl
Deutsche Forschungsanstalt fUr Luft- und Raumfahrt e.V.
Institut fUr Hochfrequenztechnik
8031 Oberpfaffenhofen
Germany
SUMMARY
For a SAR system flying on an airborne or space- flight
borne platform, the real antenna must be designed
in such a way so as to avoid ambiguities and ach- (azimuth)
ieve the envisaged resolution.
Hatched Area:
Antenna Footprint
Fig. 2 Position of resolution cell within a SAR image is determined by slant range, Doppler
frequency and time.
I flight direction
Fig. 3 Very long real aperture antenna illuminates a target The Doppler frequency shift of the
backscattered signal corresponds to the one way case.
4-3
t flight direction
0 constant frequency
/04
0./
4-
4M Doubled Doppler impact due to
two way phenomena of synthetic
O '1aperture antenna
Fig. 4 Synthetic aperture antenna illuminates a target. The signal from the real aperture antenna
propagating to the target undergoes already a Doppler shift; the backscattered signal is
"Doppler shifted" again resulting in twice the value of the one way case.
H = altitude of antenna above ground, the HPBW of the synthetic aperture
= off-nadir angle (- incidence angle),
hs = ah, * (4)
=
azimuth HPBW 3hr ahr , (2)
ah, = azimuth aperture taper factor,
X = wavelength. L = length of synthetic aperture.
The HPBW of the real aperture antenna which flies
The signal reflected from the target reaches the along the synthetic aperture (the antenna is much
different parts of the real aperture antenna with smaller than in the above mentioned comparison) is
a phase distribution corresponding to the one way used for the definition of the antenna footprint.
distance. The Doppler frequency occurring at the The length of the synthetic aperture is given by
different parts of the moving "long, real aper- the length of the footprint and therefore,
ture" antenna represent also the one way case.
).H
In the case of a synthetic aperture antenna, a L = rar = ahr 1 COS (5)
small signal source (phase centre) flies along the
synthetic aperture. We have, therefore, a two way Substituting L from (4) and Oh. from (3) in (5)
case for the phase difference along the synthetic leads to
aperture (Fig. 4), which means that phase diffe-
rences (or Doppler frequency) are twice the values ahs 1
achieved in the real aperture case of the same rap a 2 ' (6)
length. This phenomenon causes a phase change of
1800 twice as fast as in the case of the real
aperture antenna, meaning that the nulls of the which is the well known approximation for finest
synthetic aperture antenna main beam have a spa- along track resolution. For ah. = ahr, rap corres-
cing which is half of the real aperture antenna, ponds to half the length 1 of the real aperture
antenna.
Therefore, a synthetic aperture antenna has an
along track resolution (HPBW of SAR) which is ap- However, it should be noted that in many cases the
proximately twice as good as for the case of a amplitude distribution in azimuth of the real
theoretical, real aperture antenna of the same aperture antenna is not tapered, resulting in ahr
length. The along track (azimuth) resolution for a = 0.88, while the sidelobe suppression achieved by
single look image is given by the reference function of the synthetic aperture
azimuth pattern (more than 40 dB below mainlobe)
H leads to aha - 1.3. Therefore, it is a realistic
rP = c Oh. , (3) assumption to expect an rap - 0.8 1.
4-4
For zero offset processing (using in-phase and dimensions of the real aperture antenna are gi-
quadrature channels for positive and negative yen by
Doppler identification) only fD per channel is
used. Ap = h 1 > 4 ahr av H H X sino (12)
cos2o
From geometric considerations (see also Fig. I and
2), fD is defined by In many cases, ahr = 0.88 (constant amplitude)
2hr while av will be approximately 1.3 to allow sup-
fD = 2 sin - (7) pression of the sidelobes in the elevation plane.
Furthermore, range ambiguities should also not oc-
For high flying platforms (satellites), Ph is cur within the main lobe at all (main lobe width
small and is approximately 2 to 3 times larger than ov).
The distance between two radar pulses must be > Arf = 1 h . (15)
2R.,
rangetoedge
avoid the reflected
with pulse from thepulse
far
overlapping the reflected For any deviation from constant phase and ampli-
from the near range. Therefore, tude distribution, the effective antenna area will
be smaller than the real aperture. The relation-
i>2H
0 (10) ship between the two areas is called aperture
_.> 2 0826 (0 efficiency I.. It reduces the directivity proper-
tionally and therefore,
c = speed of light.
Substituting av for Ov gives the antenna height h
4-5
\e ~~
,\00..
00rI
\ cell
J~ ~~~~ i / ""
a~~~
~
v*~e~'~~C&
0 -" se tq ' ie to
L
re
q1a 12 () (18)
B = chirpin vrtial
olaizaton the are
bandwidth. til ona
The antenna gain G is smaller than the pattern di-
rectivity D due to internal antenna losses. This The ground resolution across track is accordingly
thermal efficiency, or radiation efficiency, is gvnb
defined by gvnb
For a passive microstrip antenna of several square Dual polarization capability has been increasingly
metres or a waveguide ar:" y oi ;everal metres requested in recent years in order to allow for
length (spaceborne SAR an enna , ,l might be 0.5 full polarimetric SAR operation. For instance, a
or even lower (depending on ne frequency) .horizontally polarized signal pulse will be trans-
0 " mited and the back scattered signal will be re-
ceived in the same polarization and simultaneously
o I"
these high
iConsideringwith such arrays, one solution would beassociated
one way losses to use ane ti
lower microwave
For S-band, frequencies, as L-band
suchconsiderations
and volume larization.
peated. This sequence
Assuming time be
that the will continuously
between re-
2 transmit-
and light weight
lead to microstrip arrays while for higher fro- ted pulses of different polarizations
inlpl
polarize e is
l negligi-
tra s-
oren lor (depenudigoweveefrtquency).bhorotentahlt
e fi
ch cble, we achieve the complete polarimetric matrix
array might be favoured for its lower losses
Slant range resolution ra is defined by The polarization ratio between copolarized andi
a tcross-polarized antenna pattern should exceed 30
dB within the mores in azimth and elevation. It
4,6
[km]
1. C
._o
020
--
E
cc
0 __
X100
--- 1-1A-I
0
101in
For spaceborne remote sensing, a fast off-nadir The number of T/R modules and the spacing of the
angle change capability is needed radiators (or groups of radiators) fed by a single
T/R module depends on the chosen phase controlled
- to allow scan-SAR operation, angle range.
- to increase the nuber of data take opportuni-
4-7
/ e
#.S
%Qt~1
illuminated spots
lesub
Fig. 8 Azimuth look steering mode increases the overall synthetic aperture (for multi look images
with azimuth resolution rap). Only spot-wise data takes are possible in this mode.
m~~~
so , __,
1, (20)
slm range.
2000- In
u case of At= 20* the spacing must be less than or
Kequal
-AC-B&Mmodes to 0.75 a
require Spotlight
w.beam is electronically
which or control-
azimuth look steering
Ku-Band~ Th requ
A squint
hedeired
anl orhopeainsgis
S du.1
led by phase shifters within a small (squint) angle
These modes allow high resolution observations
rrange.
_Xro nal possibilities before and after each spot (Fig. 8).
SKu-Ba The required squintptangle for such operations is
' i , -- -/ .. gousain azimut
seldom larger fedts
thanil) by
to asciiing th.
The spot
t2a. molehveao
size is fre-
- - quency dependent because of the IFOV, which increases
proportionally with the wavelength (Fig. 9).
The geeato
Equation (20) will not be applied
V atoni for squint angles
bmllaeastng Any sageri atcing fosromteom-
clof l to 2 in order to keep the number of T/R mo-
lag. I a s dules and its phase shifters small. The radiator
.... groups in azimuth fed by a single T/R module havea
ahigh directivity, which is used to suppress the
ta
h grating lobes.
i There are two causes of gain reduction due to elec-
tronic beam steering. Any angle deviating from the me-
....... (dog chanical boresight angle reduces the antenna area pro-
10 20 30 40 50 00 70 4o portionally to cosao or cosag . Furthermore, each sing-
W90 hidda, ewqb le radiator has a pattern which usually has its maxi-
mum at the (mechanical) boresight angle. The reduced
gain at other angles must be multiplied with the array
factor (which represents the far field of point sour-
Fig.9 aea or sot ighmod an/or
Ilumnatd ces arranged in the geometric layout of our antenna).
ook
teeingmodeincease
azimth apro- For instance, the gain loss of a well designed active
azimth
teeingmodeincease
ook apro- phased array at A6 = 20* will be approximately 0.5 dB.
ximately proportionally with the wavelength.
I.
4-8
6. REFERENCES
This paper is a lecture based on principles, which
have been published several times by other authors and
by recent studies, in which the author was involved.
The references below were used as sources, but due to
the general aspect of this lecture, no special refe-
rence is given in the text.
r.j ...
~5-1
A. Freeman
Imaging radar polarimetry has excited much attention The polarization of any electromagnetic wave can be
in the literature over the past few years ([-161), since characterized by the ellipticity angle, X. and the
NASA/JPL first demonstrated a successful SAR
polarimeter system in 1985 [I]. That system was known orientation angle, Vj, of the polarization ellipse, shown
as the CV-990 L-Band radar; NASA/JPL now have a fully in Figure 1. The intensity of the wave is represented by
operational, three-frequency polarimetric SAR flying the parameter Io, where
on a DC-8 platform, which has taken part in many
science data acquisition campaigns since 1988 171. 1,= al+Q (I)
Several other institutions have radar polarimeter
systems which are operational or under development, Some commonly occurring polarization states are:
including the University of Michigan, MIT/Lincoln
laboratories, the Environmental Research Institute of Vertical (linear) - (W - 00. X 00)
Michigan (ERIM) 181, the Canada Center for Remote
Sensing (CCRS). FEL-TNO in the Netherlands and the Horizontal (linear) - (4f - 900. X - 00)
Technical University of Denmark. NASA/JPL are Right-hand circular - (X 450)
currently completing the construction of a two-
frequency, fully polarimetric SAR system which will Left-hand circular- (X- 450)
fly on the Space Shuttle in 1993/1994 191. These systems
generally transmit and receive horizontally (H) and For the two circular polarization states above, the
vertically (V) linearly polarized electromagnetic fields. orientation angle. , is unspecified. Another useful way
Imaging radar polarimeters are usually implemented of characterizing
electromagnetic the
wave is polarization
the Stokes vector, state of an
using a Synthetic Aperturc Radar (SAR) approach to e
give a high resolution image in two dimensions: range
and azimuth. For each pixel in the image a polarimetric
SAR gives sufficient information to characterize the
polarimetric scattering properties of the imaged area
F =
[ I
U
V
Locos2 cos X
sin2V cos2X
1. sin
2
X
2
1o (2)
In Figure 3 the general geometry for measurements section, for any possible combination of transmit and
made by a bistatic radar is illustrated. The transmitting receive antenna polarizations (e.g. left-hand circular
antenna transmits an electric field whose components transmit and right-hand circular receive). This
are expressed in terms of a local Cartesian coordinate procedure is called polarization synthesis. For any
system (h, i, i ) with origin at the transmitting given radar receive and transmit polarization, the
antenna. We can define another Cartesian coordinate radar cross-section (RCS) can be calculated [I I via: a
system (i, y, z with origin at the scatterer. The two
coordinate systems are related via: rpq= 4X 1iv S j-P f (6)
sin u Cos i (4a) where S is the scattering matrix defined in (I). and qr
pt are polarization field vectors for the radar receive
v = -cos cos 8i - sin @lcos0 1 y + sin Oi z (4b) and transmit polarizations, respectively. For linear
polarizations (horizontal, h or vertical, v) note that the
- coOs sin Oi sin sin
Oi + os Oi (4c) RCS is given by:
{Yq )is 4x s F (7)
r Athird coordinate system (
origin at the antenna which receives
, n is defined with
the In what follows we will concentrate on linear
electromagnetic wave which is scattered by the polarizations, since most SAR systems just measure
scatterer. This coordinate system is related to the linear polarizations.
scatterer coordinate system in a similar fashion to the
above, substituting subscripts s for subscripts i in (4). Expressions (5) and (6) above represent the quantities
This choice of coordinate systems ensures that the which are directly measurable by a SAR. The units for
primed (h, ;, ) and unprimed ', v', ) coordinate Gpq are in meters squared. Both Opq and Spq are
systems coincide for a monostatic system, i.e. when the functions of spatial position I (, - ) in the scatterer
receiving and transmitting antennas are at the same
location. The radar scattering measurements made by a coordinate frame] . time (t). viewing geometry (Oj, 0
monostatic radar system are referred to as radar and radar wavelength 00) in addition to the
I7~
backwcatter measurements. polarizations of the transmitted 'and received
. ______
5-3
' Receive
Anten r--a
TransmitA
Antenna
I z -
I -/
I - - I
I a .
Scatterer
xY
Stokes matrix and Covariance matrix M 23 = 0.5 Re (SI Sh, ) - 0.5Re (Sh, S, )
representations M 24=O0.51m (Si;Sh, )- 0.51m (Sh: S,
In the monostatic scattering case, it can be shown that
the reciprocity principle [101, which dictates that M 33 = 0.5 (Sh Sh,,
) + 0.5Re (S,; S, )
I
5-4
Table I: Scattering matrices for man-made targets, normalized so that the largest
matrix element is one
Trihedral, Sphere (1
0 0)
1
PARC ilC a Ok
-sin 2a -sin a cos a)
[a is a rotation angle for each target type, about the line of sight to the radar.
a-valies of 00, 450 and 900 are commonly used, in practise.)
5-5
Circulator
H ReciverH Antenna
transmitter }P Switchi~
~Polarization
V Receiver V Antenna
Circulator
Polarimetric Calibration
(provided R R and T =T. with denoting an
Calibration of the data produced by polarimetric SAR's estimate). Note that the result still has noise
(e.g. 111. 1161, 117]) is necessary if the data is to be used superimposed on it. At the complex SAR image stage
in a quantitative fashion. Goals for calibrating (which is what polarimetric SAR scattering matrix data
polarimetric SAR data are given in Table 2. The basic is), although the mean noise contribution is zero, the
problem in calibrating polarimetric radar data can be rms uncertainties due to noise can not be removed.
seen from the following expression, after (181. for the rsucranisdet os a o ermvd
meeasuremets e byllowg
axporeric, ar .forwhh After square-law detection, however, the average noise
measurements made by a polarimetric radar for which poe ca bestrtdofth rsuin RC
R ad
ysem Tmatixode isvald:power can be subtracted off the resulting RCS
the 2x2 R and T matrix system model isvalid: measurements (though this still does not remove the
rms uncertainties due to noise). Polarization synthesis.
Ma,'Wei' Shh Svh 16 in which the target RCS, Gpq, in response to a radar
with transmit polarization p and receive polarization q
82 fl Sh Sw 84 f 2 is synthesized from the scattering matrix, involves a
square-law detection step, so the noise bias term can be
rnhh nlh subtracted off the synthesized image pixel powers,
I nhv
+ v ~n( nvv J provided it can be estimated.
We characterize the (complex) noise terms as having
nN
V=K/ e*,Rt ST + v/- (13) two-dimensional, zero-mean, Gaussian distributions.
with the following properties:
The validity of the majority of the polarimetric the problem, we need at least another six equations in
calibration algorithms in the literature depends on the the various terms involved. There are three approaches
validity of the system model above for radar to obtaining the required six equations: the first is to
polarimeters, which was first put forward in [I]. This use man-made targets with known scattering matrices;
system model contains just six relative parameters, the second is to make assumptions regarding the
including four cross-talk (or leakage) terms and two general properties of the backscatter being measured:
channel imbalance terms, one for the H,V transmitters and the third is to make assumptions about the radar
and one for the H,V receivers. Determination of these system parameters. Of course, all three of these
six parameters, followed by correction for any approaches may be combined to find an appropriate
deviations from the ideal is then sufficient to calibrate solution.
the radar data polarimetrically, so that the HH, HV, VH
and VV scattering matrix measurements can be The first attempts ([201 - [221) to calibrate polarimetric
meaningfully compared. radar systems used only the man-made target
approach. There are only a limited set of such targets
The generally accepted polarimetric system model available. Table I lists the normalized scattering
relies on the constancy of the system, particularly the matrices for some of the more common ones. Barnes
transmitters, the receivers and the antennas. For 1211 and Woods 1221 both used combinations of three
instance, it assumes that the ratio of the power passive devices to come up with the six additional
transmitted between H and V polarizations never equations required to solve the problem. In his
varies, and that the receiver gains for H and V are kept approach, Woods 1221 used a trihedral and two dihedrals,
.
one at 00 rotation, the other at 450 Barnes 1211 also used
at a constant level relative to each other. This is not this approach, and presented another solution using
always the case for radar polarimeters [191. For two dipoles (at 00 and 900 rotation angles) and one 450
example, in operating the NASA/JPL system the gain of dihedral. The dipole signatures were achieved using
the H (or V) receiver can be switched by up to 6dB, screened trihedral corner reflectors ([231, [241). Three
depending on whether a like-polarized (HH or VV) or targets were found to be necessary to solve the problem
cross-polarized (VH or HV) return is expected. [ As because it was very difficult to obtain more than two of
described above, the NASA/JPL system has two the required equations from each target signature.
receivers, one for H and one for V, and alternately Tueh et al [251 presented a more general solution using
transmits H and V pulses!. The reason for operating the passive targets, in which at least two must either have
radar this way is that the cross-polarized returns from singular or non-singular scattering matrices, and the
natural targets are usually lower than the like- targets must not be the same. Whitt and Ulaby [261
polarized returns, so the gain is increased when showed how the problem may be solved provided just
expecting a cross-polarized return to keep the signal one of the three target scattering matrices is non-
level within the optimum range for the analog-to- singular. Freeman, Werner and Shen [151 showed how
digital converters. This gain differential should be three polarimetric active radar calibrator (PARC)
removed during processing. The ERIM P-3 polarimetric signatures could be used to solve the problem. The
SAR solves the same problem with a different design, sensitivity of all of these approaches to small errors in
havigtansmtte,
oly oe on reeive an an the rotation angles of the devices used was pointed out
automatic gain control (AGC), which varies the gain of in [191.
the radar receiver in some pre-set fashion [8]. The AGC
setting in amplitude and phase is usually different for
HH, HV, VH and VV returns, which are collected in and the backscatter to arrive at a solution to the
sequence. The design for the CCRS polarimetric SAR has
made the following assumptions
Zyl [27! radar
an elaborate switching sequence for it's two receivers, about the van
problem. polarimetric system:
so that HH and VV returns go through the same
receiver, and VH and HV returns go through the other
receiver. In none of these cases is the 2x2 R and T 61 =
matrix system model necessarily valid, unless the gain 62 864
differentials between receivers are properly removed
from the data. fI f2 (17)
Let us now return to the problem of estimating R and T, i.e. R - T. These are three of the required equations.
in order to calibrate the scattering matrix data. van Zyl also assumed that the backscatter was
Expanding (13), we obtain four equations relating the reciprocal, i.e. Shy - Svh, which would at first seem to
scattering matrix measurements with their actual provide a fourth equation, but on further examination
values: of equations (16b) and (16c) this makes one of them
redundant. so the net number of equations does not
Mhh - Shh + 8 2Shv + 84Svh increase. Next, van Zyl assumed that for most natural
++ nhh (16a) backscatter types, azimuthal symmetry holds [3]. for
8284vv hhwhich the like- and cross-polarized backscatter terms
are uncorrelated, i.e.,
+
Mhv- 81Shh fIShv + 8 184Svh (16b)
+ fl84Svv + nhv (Shh Shv* 0
8 + 8 8 +
Mvh - 3Shh 2 3Shv f2Svh
(0hvSIv 0
+ f282Svv + nvh
This gave two more equations, which allow the
8 8
Mvv - 1 3Shh + fl83Shv + f28lSvh (16d) calculation of the two cross-talk terms, 81 and 81. These
+
+ flf2Svv nvv can vary significantly with incidence angle as can be
seen from the plots shown In Figure 9. The solution was
completed by estimating flf2 from the ratio of the HH
These four equations contain ten unknow i quantities and VV measurements from a trihedral. From (16a) and
(ignoring the noise), which are the four (complex) (16d) It Is easy to see that for a trihedral, with
scattering matrix elements and six (complex) radar scattering matrix as given in Table|,
system unknowns (the 8's and the f's). Clearly, to solve
5-8
140
13 00
-20a 60-
" 20
-30 ' 2
-40 -60-
-loo
; -:4 n~ delta
,m~m
mm lmmlln Il i i
mallmmm
5"
5-9
assumptions that the different pairs of cross-talk terms is short (less than 10 meters), densely packed, and has
were equal (but not that f I - f2). This was followed by a sparse foliage. Yet another wetland forest type is
rotation of the H-V coordinate system such that 81 - 82, swamp forest, consisting of short trees or shrubs
with a trihedral signature then yielding the remaining growing in standing water. Besides forest, there are
three equations. reed and sedge marshes. These are areas of herbaceous
vegetation in standing water. The sedges are
Calibration of the phase difference between considerably taller than reeds (200 cm versus 40 cm).
polarization channels (i.e. determination of arg(fl) and In addition to these naturally occurring vegetation
arg(f2) alone) using corner reflector signatures and types there are agricultural areas consisting of cleared
areas, bare soil and various crops.
signatures of moderately rough surfaces is addressed in
1341. An examination of the assumptions on the system
and backscatter behavior required for this approach to Table 3 contains typical radar backscatter
be successful was presented in (351. measurements for the different vegetation types
identified in the area imaged by the DC-8 SAR. From the
Polarization Filtering table it can be seen that the very low radar cross
sections correspond to open water and bare ground at
We have seen that it is possible to synthesize the all three frequencies. The upland forest spans a fairly
response of any target to arbitrary combinations of narrow range of cross-sections. The range of radar
transmit and receive polarizations from measurements cross sections is given in the table by the two values
of the scattering matrix, the covariance matrix and the shown in the table. Radar cross section tends to
Stokes matrix. We have also seen from the example increase as the angle of incidence decreases. If
polarization signatures that transmit/polarization incidence angle effects are considered, the radar cross
combinations exist for which the synthesized radar section for the upland forest is essentially constant. For
cross-section is maximized or minimized and that the other types of vegetation, at P-Band we see a steady
particular combinations of transmit and receive climb in the HH cross section as the biomass of the
polarizations for which the maximum and minimum canopy increases until we reach a plateau where the
RCS values are reached will depend on the type of upland forest begins. The same feature is visible in the
scattering (e.g. Bragg, dihedral mechanism). Consider L-Band and C-Band data, but at L-Band the plateau is
the problem of trying to detect an isolated target (e.g. a reached for re-growth vegetation (2-3m in height).
ship) against a clutter background (e.g. the ocean) while at C-Band the plateau is reached for farmland
using a polarimetric SAR. In this case we would choose vegetation (<in in height). This type of saturation of
the transmit and receive polarization combination radar backscatter with biomass has been observed
which maximized the ratio of the synthesized target RCS previously. At all three frequencies, especially C-band,
over the average of the synthesized clutter RCS. i.e. the we then see a secondary increase region in the HH
signal-to-clutter ratio: radar cross section data. The sites corresponding to
these data points may have much less biomass than the
lI ) upland forest, but they have a brighter radar return.
t some feature which adds a
SCR . - q(aVr, Xr) ST P(Wt, X) (20) These areas generally have
secondary scattering mechanism in addition to the
q(vr,t
Xr~ Sc p(. At". Ivolume scatter which dominates the upland forest. One
site was a rain forest area where the trees had been cut
SCR is also known as the contrast ratio. Several authors but not cleared (clear-cut in Table 3). The site had an
have offered mathematical solutions to this problem enhanced HH polarization return due to the many tree
([361-1381). The first step is to characterize the trunks laying on the ground. Other sites with
scattering properties of the target and the clutter, i.e. on
to enhanced backscatter were sites with vegetation
growing in standing water (reeds. sedge, high marsh
determine ST and SC. Then the polarizations
transmit and receive which maximize the SCR have to forest, swamp forest, flooded Bajo). In this case the
be determined. Application of those transmit and backscatter probably consists of volume scatter from
receive polarizations to an entire image is known as the vegetation canopy plus double bounce scattering
polarization filtering, due to interaction between the vegetation and standing
water. In other sites, the double-bounce mechanism
Radar mapping of vegetation types may be enhanced because the vegetation understory is
relatively clear (in the areas designated coffee and
Due to the concern in the scientific community over palm forest, for example).
the global warming problem. mapping of vegetation
cover (and changes in vegetation cover), especially in A simple, three-component scattering model
tropical rain forest areas, using remote sensing data
has received considerable attention recently. Synthetic We can begin to understand the phenomena which give
Aperture Radar, with it's ability to collect data during rise to the radar backscatter characteristics seen in
day or night. or in cloudy conditions, is ideally suited Table 3 and in other polarimetric SAR images of natural
for monitoring the earth's remaining tropical rain surfaces by using a very simple conceptual model of
forests. scattering. This model is more fully developed in [391;
here an outline is presented. We assume that scattering
In what follows, we present multi-frequency. from a veRetation laver is a combination of volume
polarimetric SAR results from a tropical rain forest site scattering from vegetation, double bounce scattering
in Belize, Central America, as an example of the ability from the ground/trunk interaction and scattering from
of polarimetric SAR to map different vegetation types. a rough surface. For volume scattering, we assume that
An image of a 10lOkm area is shown in Figure 10. The the return is from randomly oriented, very thin
area under study is dominated by old growth upland cylinder-like scatterers. The double bounce scattering
tropical rain forest. There are also small areas which component can be modeled by scattering from an
have been cleared of upland forest within the last five upright (rotation angle - 00) dihedral corner reflector .
years and allowed to re-grow. In addition, there are In the model, both surfaces in the reflector are made of
areas of almost pure palm forest, consisting of either dielectric material, and the two surfaces may have
cohune or botan palm trees. There are several types of different dielectric constants, corresponding to trunk
wetland forests. High marsh forest has vegetation and ground. The trunk has Fresnel reflection
wateris coefficients and Rtv forLikewise. the and
horizontal vertical
resembling the upland
is present during
rainportion
a large thestanding
forest,ofbut year. Bajo polarization, Rth
respectively. gr6und has
another type of wetland forest which grows in areas Fresnel reflection coefficients Rsh and R3 p. For the
having very poorly drained clay soils. The vegetation
~5-10
Vegetation -and D -and
0
Class OHH HVV P HH/VV HV/HH (JHH IHHVV P HH/VV HV/HH HH 44HVV P HH/VV /U
Open Water -32.6 5.8 033 -6.6 -7.4 -28.7 -3.1 0.63 -7.7 -6.6 -23.7 -18.7 029 -2.4 -5.3
Bare Soil -25.1 -8.8 0.75 -5.4 -9.5 -16.5 -23.7 0.75 -1.8 -10.4 -9.0 -10.0 0.76 -0.3 -8.9
Reeds -23.0 115.3 0.72 -10.9 -10.7 -9.0 98.1 0.54 -3.3 -11.6 -0.3 135.8 0.44 22 -11.9
Farmland -20.3 -12.5 0.53 -2.0 -11.5 -13-3 -18.6 0.75 -1.5 -11.7 -7.6 -12 0.47 0.7 -6.4
Regrowth -15.3 20.7 025 -0.9 -5.8 -10.1 -22.7 0.31 -0.1 -5.5 -5.7 3.2 0.44 0.7 -6.5
Bajo -14.5 70.8 021 -2.8 -5.8 -9.7 15.1 0.17 0.1 -5.6 -6.2 -2.1 0.42 -0.2 -5.6
Swamp Forest -13.8 149.5 0.1 -0.6 -8.4 -6.9 165.4 0.06 0.4 -7.6 -2.1 2.3 0.41 1.2 -8.9
Upland Forest -12.8 32.3 022 -0.8 -5.4 -9.2 7.9 0.25 0.2 -5.1 -7.4 9.0 0.46 0.0 -5.8
Upland Forest -11.5 51.1 0.14 0.4 -6.4 -9.2 7.9 0.25 0.2 -5.1 -5.6 -4.1 0.51 0.6 -6.6
High M. F. 2. -11.0 6.4 0.16 -2.3 -5.3 -8.4 7.8 0.16 0.2 -5.2 -5.3 -1.4 0.43 0.6 -5.8
Palm Forest -11.3 91.4 021 -0.1 -7.2 -8.6 48.5 0.2 0.6 -6.4 -5.0 -2.4 0.56 0.4 -7.0
Sedge -10.3 93.6 0.18 0.5 -7.0 -8.8 -5.7 0.26 -0.2 -5.2 -6.1 -0.9 0.52 0.0 -6.3
Flooded Bald -9.7 75.6 0.16 0.9 -7.0 -8.6 31.4 0.22 0.6 -5.6 -6.4 -2.3 0.42 -0.9 -5.5
Coffee -9.2 137.3 0.4 1.1 -8.8 -8.0 52.1 0.12 1.7 -7.7 -6.6 6.9 0.42 1.4 -7.0
Clear-cut -9.0 31.0 0.32 3.0 -8.5 -7.6 6.9 0.4 2.0 -8.0 -6.0 8.3 0.55 1.4 -8.3
High M. F. 1 -8.2 -32.7 02.9 -1.6 -7.5 -4.0 -2.5 0.32 2.1 -7.8 -3.2 15.0 0.35 1.5 -7.0
" Cross section results and ratios are in dB; phase differences are in degrees
* High M. F. - High Marsh Forest
Open Water 0.67 0.09 0.24 0.76 0.01 0.23 024 0.05 0.70
Bare Soil 0.77 0.03 0.20 0.65 0.07 0.27 0.56 0.01 0.44
Reeds 0.02 0.93 0.05 0.16 0.67 0.17 0.11 0.59 0.30
Farmland 0.65 0.16 0.19 0.69 0.08 0.23 020 0.01 0.79
Regrowth 0.14 0.09 0.76 0.08 0.05 0.87 0.19 0.02 0.78
Bajo 0.10 0.26 0.64 0.02 0.11 0.87 0.14 0.01 0.85
Swamp Forest 0.16 0.37 0.47 0.09 029 0.62 036 0.12 0.52
Upland Forest 0.04 0.12 0.84 O.02 0.03 0.95 0.16 0.01 0.83
Upland Forest 0.05 0.18 0.77 0.00 0.05 0.95 023 0.01 0.76
High M. F. 2 0.04 0.23 0.73 0.00 0.03 0.97 0.12 0.00 0.88
Palm Forest 0.02 0.39 0.59 0.08 0.14 0.78 029 0.01 0.70
Sedge 0.07 0.25 0.68 0.02 0.06 0.91 024 0.01 0.75
Flooded Bajo 0.05 0.21 0.73 0.00 0.00 0.94 0.17 0.02 0.81
Coffee 0.00 0.46 0.54 0.07 0.25 0.69 0.23 0.04 0.74
Clear-cut 0.30 0.07 0.63 0.29 0.06 0.65 0.37 0.04 0.59
High M. F. I 0.30 0.19 0.51 0.24 0.07 0.69 0.19 0.05 0.76
Table 4: Backscatter mechanisms from the model fit for typical examples of the
different vegetation classes.
surface scatter, we assume a first-order Bragg model where fs. fd and fv are the surface, double-bounce and
(12) is adequate to describe the backscatter. For all of volume scatter contributions to the VV cross section, P
these three backscatter components, we assume that
like- and cross-polarized returns are uncorrelated. and is a real number and Eis given by:
that the backscatter is reciprocal (HV - VH). Now, if the
volume, double-bounce and surface scatter components Q _Rh R
are uncorrelated. the total second order statistics are a Rg Rt
the sum of the statistics for the individual mechanisms; Rgv Rtv (25)
i.e., the backscattered powers rather than voltages are
added. Thus our model for the total backscatter is: This model gives us four equations in five unknowns.
In general, a solution can be found if one of the
unknowns is fixed. Since neither the surface or double-
bounce mechanisms contribute to the HV term in the
h 2 fs 1 fd Gf (21) model, we can use this to estimate the volume scatter
contribution directly. The volume contribution can
2 then be subtracted off the IShh 12. ISvv 12 and ShhSvv*
IS~Ic
j " 1 +
fd +
fv (22) terms, leaving three equations in four unknowns. After
1401, we then decide whether double-bounce or surface
3
,-. fv/ (23) scatter is the dominant contribution in the residual
based on the sign of the real part of ShJSvv" . If Re(
ShhSvv" ) is positive, we decide that surface scatter is
ShbS,v - P fs " a fd fv/ 3
(24) dominant, and fix a - 1. If Re( SbSvv" ) is negative, we
J
r ~5-l11
decide that double-bounce scatter is dominant in the performing supervised classification on polarimetric
remainder and fix 0 - I. Then we calculate fs, fd and 0 radar data. First training sets are selected, which
or Ot from the backscatter measurements. In Table 4 we represent different surface types. Then the elements of
the covariance matrix for each surface type can be
show the fraction of the total power contributed by averaged to find the average polarimetric scattering
each of the three mechanisms (with odd - surface, dbl properties for each class. Kong et al (411 developed a
double-bounce and vol - volume scatter) for the supervised classifier in which the following vector is
vegetation classes given in Table 3. formed for each pixel in the image:
Another type of classifier is known as the minimum 181 Sullivan, R. et al, Polarimetric XIL/C-band SAR,
distance classifier. For polarimetric SAR data, a Proc. IEEE National Radar Conf., Ann Arbor. MI,
minimum distance classifier was implemented by van pp. 9-14, 1988
Zyl and Burnett@. For this type of classifier the distance
measure used is: 191 Carver, K.. et al, Shuttle Imaging Radar-C Science
Plan, JPL Publication '86-29, 1986.
(101 Zebker, H. A. and van Zyl, J. J.. Imaging Radar
Polarimetry: A Review, Proc. IEEE, vol. 79. No. 11.
where November 1991, pp. 1583-1606.
[Il 1 Ulaby, F.T. and Elachi, C. (ed.), Radar Polarimetry
for Geoscience Applications, Artech House. 1990.
S,, SI:
(121 BoernerW. M.. et al (eds.), Inverse Methods in
- S Electromagnetic Scattering, Hingham. MA:
S SZJ (31) Reidel. 1985.
SS[131 van Zyl, J. J., Zebker, H. A. and Elachi. C.. Imaging
S& Sf; radar polarization signatures: theory and
observation, Radio Sci.. vol. 22. pp. 529-543. 1987.
[141 Dubois. P.C. and Norikane, L., Data volume
i.e. a subset of the covariance matrix elements and the reduction for imaging radar polarimetry, Proc.
decision rule is that the scatterers which are IGARSS '87, Ann Arbor. MI. pp.691- 696, 1987.
represented by Y are a member of the class i ir [15] Freeman, A., Shen. Y. and Werner, C.L,,
Polarimetric SAR Calibration Experiment Using
d, nY< d nY-forallj *i (32) Active Radar Calibrators, IEEE Trans. on
Geoscience and Remote Sensing, Vol. GE-28. No. 2.
March 1990.
Acknowledament 1161 Freeman. A., Werner, C. and Shen. Y., Calibration
Part of the work described in these notes was carried of Multipolarization Imaging Radar. Proc.
out by the Jet Propulsion Laboratory. California IGARSS'88, pp. 335-339. 1988.
Institute of Technology, under a contract with the
National Aeronautics and Space Administration. 1171 Freeman, A., Shen. Y., van Zyl. J.J. and Klein. J.D..
Calibration of NASA/JPL DC-8 SAR data. Proc.
IGARSS '9 1, Espoo, Finland, June 199 1.
Bibliography 1181 Freeman. A.. SAR Calibration: An Overview.
submitted to IEEE Trans. on Geoscience and
IlI Zebker, HA.. van Zyl, J.J. and Held, D.N.. Imaging Remote Sensing. October 1991.
Radar Polarimetry From Wave Synthesis. J.
Geophys. Research, Vol. 92. pp 683-701. 1987. 1191 Freeman, A.. A New System Model for Radar
Polarimeters, IEEE Trans. on Geoscience and
[21 van Zyl, J.J., Papas. C.H. and Elachi, C., On the Remote Sensing, Vol. 29, No. 5, pp. 761-767,
optimum polarizations of incoherently reflected September 1991.
waves, IEEE Trans. on Antennas and Propagation,
vol. AP-35, pp. 8 18-825, 1987. [201 Livingston, P.S. and Kaplan, D.J., Radar
Calibration Procedure, Technical Note 1432,
131 Borgeaud, M.. Shin, R.T. and Kong, J.A.. Naval Ocean Systems Center, San Diego, CA,
Theoretical models for polarimetric radar clutter. November 1985.
J. Electomagnetic Waves and Applications, vol. 1.
pp. 67-86, 1987. 1211 Barnes, R.M.. Antenna Polarization Calibration
Using In-Scene Reflectors, Lincoln Laboratory
[41 Boerner, W. M., Foo, B.-Y. and Eom, H. J.. Technical Report TT-65. Sept. 1986.
Interpretation of polarimetric Copolarization
phase term in the radar images obtained by the [221 Woods. M.A., A. Calibration Procedure For A
JPL airborne L-band SAR system, IEEE Trans. on Coherent Scattering Matrix Radar. Royal Signals
Geoscience and Remote Sensing. vol. GE-25, pp. and Radar Establishment. Memorandum 3889.
77-82. 1987. May 1986.
15[ Ulaby. F.T., Held, D.N., Dobson. M.C.. McDonald, K. (231 Kennaugh, E.M., Polarization Properties of
C. and Senior, T. B. A., Relating Polarization phase Corner Reflectors With Modified Walls, Ohio State
Difference of SAR signals to scene properties. University, Report 612-6. January 1957.
IEEE Trans. on Geoscience and Remote Sensing,
vol. GE-25, pp. 83-92. 1987. 1241 Kennaugh. E.M. and Chang, C., Design and
Theoretical Performance of Reactive-Wall
161 Evans, D. L., Farr, T. G., van Zyl, J. J. and Zebker, Corner Reflectors, Ohio State University
H. A., Imaging radar polarimetry: analysis tools Technical Report NOW 64-0215-D January 1965.
and applications. IEEE Trans. on Geoscience and
Remote Sensing. vol. GE-26, pp. 774-789, 1988. (251 Yueh, H.A., Kong, J.A.. Barnes, R.M., and Shin,
R.T., Calibration of Polarimetric Radars Using In-
(71 Held, D.N., at al, The NASA/JPL multifrequency, Scene Reflectors, J. of Electromagnetic Waves &
multipolarizatlon airborne SAR system, Proc. Applications..
IGARSS '88, Edinburgh, Scotland, pp. 345-349.
1988.
I.
5-13
[261 Whitt, M.W., Ulaby. F.T., Polatin. P. and Liepa, (341 Zebker. H.A., and Lou, Y.. Phase Calibration of
V.V.. A General Polarimetric Radar Calibration Imaging Radar Polarimetric Stokes Matrices. IEEE
Technique, IEEE Trans. on Antennas and Trans. on Geoscience and Remote Sensing, Vol.
Propagation, Vol. AP-39, No.1, pp. 62-67, January GE-28, No. 2, pp. 246-252, March 1990.
1991.
[351 Sheen, D.R., Freeman. A. and Kasischke, E.S..
[271 van Zyl, J.J., Calibration of Polarimetric Radar Phase Calibration of Polarimetric Radar Images,
Images Using Only Image Parameters and IEEE Trans. On Geoscience and Remote Sensing,
Trihedral Corner Reflector Responses, IEEE Vol. GE-27, pp. 719-731, Nov. 1989.
Trans. on Geoscience and Remote Sensing. Vol.
GE-28, No.3. pp. 337-348, May 1990. [36) loannidids, G. A. and Hammers. D. E.. Optimum
Polarizations for Target Discrimination in
[281 Klein, J.D.. Calibration of Complex Clutter, IEEE Trans. on Antennas and
Quadpolarization SAR Images Using Backscatter Propagation, Vol. AP-27. pp. 357-363. 1979.
Correlations, submitted to IEEE Trans. on
Aerospace and Electronic Systems, 1989. [371 Dubois, P. C. and van Zyl, J. J., Polarization
filtering of SAR data, Proc. IGARSS '89, Vol. 3. pp.
1291 Freeman, A., van Zyl. J.j., Klein. J.D.. Zebker. H.A. 1816-1819. 1989.
and Shen, Y., Calibration of Stokes and scattering
matrix format polarimetric SAR data, accepted [381 Swartz, A. A., Yueh, A. H., Kong, J. A.. Novak. L. M.
for publication in IEEE Trans. on Geoscience and and Shin. R. T., Optimal polarizations for
Remote Sensing. 1991. achieving maximum contrast in radar images.
Journal of Geophysical Research, Vol. 93, No. B12,
[301 Klein. J.D. and Freeman. A., Quadpolarization SAR pp. 15252-15260. 1988.
Calibration Using Target Reciprocity. Journal of
Electromagnetic Waves & Applications, Vol. 5. No. [391 Durden. S.. et al. Multi-frequency polarimetric
7. pp. 735-751. 1991. radar observations of a tropical rain forest.
submitted to IEEE Trans. On Geoscience and
[311 Sarabandi. K. and Ulaby, F.T.. and Tassoudji, M.A.. Remote Sensing, April 1992.
Calibration of Polarimetric Radar Systems With
Good Polarization Isolation, IEEE Trans. on [401 van Zyl, J. J., Unsupervised classification of
Geoscience and Remote Sensing, Vol. 28, pp. 337- scattering behavior using radar polarimetry
348. May 1990. data. IEEE Trans. On Geoscience and Remote
Sensing. vol. 27. pp. 36-45, 1989.
[321 Whitt, M. and Ulaby. F.T.. A polarimetric
calibration technique with insensivity to target [411 Kong. J. A.. Swartz. A. A.. Yueh, H. A.. Novak. L. M.
orientation, Proc. IGARSS '90, Washington, D.C.. and Shin. R. T., Identification of Terrain Cover
pp. 1089-1092, May 1990. using the Optimum Polarimetric Classifier.
Journal of Electromagnetic Waves and
[331 Sarabandi, K.and Ulaby, F.T., A Convenient Applications. Vol. 2. pp. 171-294. 1988.
Technique For Polarimetric Calibration of
Single-Antenna Radar Systems, IEEE Trans. on
Geoscience and Remote Sensing. Vol. GE-28, No. 6.
pp. 1022-1033. November 1990,
4)
6-I
Hydrologists, especially, requested an absolute Th'nmain emphasis of the lecture will be on the
calibration with tolerances of less than 1 dB. In- spaceborne sensors, because an angle dependence of
ternal calibration schemes, as well as the use of a' has a strong impact on image evaluation due to
external passive and active calibrators, were in- the large incidence angle range coverage of air-
troduced to achieve this goal over the wide dyna- borne systems.
mic range. The in-flight measurement of the anten-
na pattern by means of ground based receivers be- 3. RADIOMETRIC CORRECTION OF NEAR RANGE/FAR RANGE
came increasingly important due to pattern changes DIFFERENCES AND OF ANTENNA PATTERN INFLUENCE
caused by electronic beam steering and the neces-
sity of beam alignment in case of multifrequency Spaceborne SAR sensors usually have a half power
and/or polarimetric operation modes. beamwidth (HPBW) of a few degrees in elevation.
Therefore, the influence of incidence angle chan-
The use of radiometric corrections to compensate ges on 00 can, in many cases, be neglected if the
for near range/far range differences caused by an- instantaneous field of view (IFOV) covers flat
tenna pattern and geometry of illumination will be terrain. The range dependence cannot be neglected,
explained in the lecture. because it affects the signal-to-noise ratio S. by
a power of 3.
The impact of geocoding on radiometric levels will
also be mentioned, including the phenomenon of A similar statement can be made for the antenna
over-/underexposing hilly regions, caused by inci- gain directional pattern G(6), since the signal-
dence angle changes. to-noise ratio increases with the square of G(b),
as can be seen in the simplified SAR-equation be-
2. INTRODUCTION low:
2
For the Radarlab missions, SIR-C/X-SAR Science
Plans [1, 2] were published which described expe- S- '16) K , (1)
riment goals within the various disciplines (such R3 (6) sin6
as geology, hydrology, glaciology, vegetation, Pav 13 ahs rR
technology). At that time, it was not possible to K =
guarantee SAR system performance as well as was 20W kTo F u as
requested by the scientific user community (se- 2 T
lected principal investigators and co-investiga- K is assumed to be constant, because the average
tors). Critical parameters were instrument stabi- p is wvelent zimu
t, te acerage
lity and relative and absolute radiometric cali- power Pay, wavelength , azimuth taper factor a,,
bration. Consequently, a number of selected expe- (of the synthetic aperture), slant range resoluti-
field of technology
in the calibration. are dedicated on Boltzmann's
rR , figure constant k, temperature T.,
riments
to radiometric The engineers des- noise F, platform velocity u and pulse ta-
igning and manufacturing the SAR sensors were re- per factor an do not change with respect to the
quested to provide internal calibration loops. off-nadir angle 0.
Means for X-SAR internal calibration are described
in [3]. Similar loops
• were integrated in the ERS-1 In orderassuming
swath, to keep atS() seeming
the same timeconstant across
that 00(0) is the
[4]. constant across the swath, the following correc-
4
An overview of the efforts and plans of f ve years tion function CF is required:
ago was published in [5], and one year ago a
workshop was dedicated to SAR calibration [6]. CF = R3 () sino (2)
taking into account viewing angles, spacing and 40" off-nadii (Fig. 1).
6-2
Antenna
i4
, Nadir
SI
0I range
6-
0
-3.0 -1.5 0.0 1.5 3.0
off-boresight angle [dog]
Fig. 2 Correction function CF across the swath. Near range is on the left, far range on the
right; boresight angle of antenna is 00.
The computed correction function CF across the 4. IFLUZXM OrU3ORZtIGT MRS DUR TO U
swath within the HPBW is shown in Fig. 2. The ATTITUDE ZRO IN AOLL DIRECTIN
asy metric shape is caused by the illumination
geometry. For known antenna pointing and given (or measured)
pattern shape inelevation, it ispossible to ap-
m ODA
~6-3
I
-10-
c-20
-30
-40
-10 -5 0 5 10
off-boresight angle [deg]
Fig. 3 Predetermined X-SAR elevation pattern.
2 I I I I I I I I I
11
L.%
-2,
-5 4 3 2 1 1 2 3 4 5
off-boresight angle [deg]
Fig. 4 Gain error function for various boresight angle errors; solid line ± 0.3*, and dashed-
dotted line ± 0.40.
ply equation (2)to correct the image amplitude. 5. ANTRNNA PATTERN Z4ZSUENZNT
In case the pointing error is unknown, a radiome- As mentioned in part 3 of this paper, the antenna
tric error will be produced which, in most cases, pattern must be known for precise radiometric cor-
cannot be corrected. rection. In view of strong mechanical stress du-
fne image).
the prpecise
the measured azimuth-cuts can later be time-cor-
related and the required elevation main cut of the
antenna pattern can be reconstructed. Of course,
location of the SAR-sensor vs. time
must be known to convert the time of pulse receipt
6-4
4 !
-Elevati-
Schematic m
160 1 1 V .
120-
00 80- H H
0
40-
0 15 0 15
SAMPLE NUMBER
Fig. 6: Registered pulseshapes from NASA/JPL's C-band RHDC-8 SAR.
into antenna angles. It is also mandatory to pre- The relatively large deviations of the measured
cisely know the attitude data of the platform as points from the fitted curve in Fig. 7 can be ex-
well as the location of each receiver. plained by 3 facts:
Inthe 1991 NASA/JPL airborne SAR-campaign over 1)precise flight and attitude data have not yet
our test site, we had a first opportunity to test been received from,
ourr m2)
measuring system (9].
the positions of the receivers were only known
Fig. 6 shows registered radar pulses from the to about t 100m,
C-band SAt of NASA/JPL's DC-8 (horizontal po-
larization). Slight distortions, i.e. decreasing 3) at the time of the experiment not all of our
power from begin to end of the pulse can be seen, equipment was fully calibrated.
probably due to the wide chirp used and the rela-
tively small antenna bandwidth. These drawbacks will be eliminated in our future
experiments with spaceborne SAR-sensors (all re-
An azimuth cut can be obtained, by integrating the ceivers are fully calibrated now, positions are
pulse energies andofplotting themcuts
against
givestime.
time-correlation all azimuth the The determined using differential
bit and attitude GPS andfrom
data are available precise or-
the sa-
main cut of the antenna elevation pattern (see tellite owners).
Fig. 7).
The co-registration of the time of pulse receipt
.1
6-5
-5 A
C
C
-0 _ ___
.N
ca
M
E -15
0
0 15 30 45 60 75 90
off-nadir angle [deg]
Fig. 7 Reconstructed main cut of the antenna elevation pattern.
.?l7~ -02dB 1 I 1
HPA
'"-d~.-
FGU
can yield two sets of further information: fore and after data takes. It can be used to dedu-
ce an optimum reference function for processing.
Precise measurement of cross-band beam alignment The test signal can either be taken from the low
in azimuth and elevation direction (especially power part of the radar transmitter (HPA is swit-
suitable in the X-SAR/SIR-C mission) and possible ched off) or from the directional coupler between
antenna squint angles. the HPA and antenna. In the latter case, the low
noise amplifier (LNA) will be bypassed. These
Our system will be used in all present and near- loops are indicated in the block diagram in Fig.8.
future SAR missions (ERS-l, JERS-l, X-SAR/SIR-C,
PRIRODA). Internal calibration of active array antennae is
more complicated due to the large number of trans-
6. INTERNAL CALIBRATION mit/receive modules (T/R modules), which can easi-
ly reach several hundred to several thousand.
Spaceborne SAR systems should have life times be-
tween 3 and 6 years without too much degradation These T/R modules f the next generation SAR an-
in performance. Internal calibration loops are ad- tennae will be designed and manufactured in mono-
visable to monitor system stability. For the high lithic microwave integrated circuitry (wIC)ba-
power output, monitoring can be achieved by measu- sed on GaAs technology. Besides control and moni-
ring a small portion; the measuring point could be tor units (power, phase shifter, polarization), a
a directional coupler be..ween the high power am- time sequential check-out of the transmitter and
plifier (HPA) and the antenna. receiver stages is necessary. Different concepts
using separate feeding networks to distribute or
Receiver gain stability and linearity can be mea- collect calibration signals within the active ar-
sured by feeding a small part of the radar signal rays are presently under study at various centres.
via attenuators through the whole receiver chain.
For power leakage problems, the HPX is usually For internal calibration, the balance and orthogo-
switched off during this test phase and the signal nality of in-phase and quadrature channel (after
is taken from a low power section. demodulation) are important. Error analyses for
X-SAR have shown that 50 deviation from orthogona-
Furthermore, the actual chirp can be measured be- lity causes about 0.2 dB amplitude error and 0.2
I 1i .
6-6
40
30 _
20
transfer function
__ 10 - a comer reflectors
TRACK
2
-10 I--grass C - Band / VV
-20 rconcrete col.-factor: (18.6*0.6) dB
-30 1 m 1 , 1
0.0 100 200 300 400 500 600
image amplitude
Fig. 9 E-SAR system transfer function.
30
20
10 - transfer function
6 a corner reflectors
b 0-0
-10 TRACK 2
gross C - Band / VV
-20 concrete col.-factor: (6.1 *0.8) dB "
-30 1 1
0.0 25 50 75 100 125 150
image amplitude
Fig. 10 DC-8 SAR system transfer function.
Another source of SAR signal contamination is the External calibration can only be usefully perfor-
total system noise. This noise is mainly composed med after having accomplished the previously des-
of contributions from the sensor itself (ADC, LNA, cribed internal calibration and radiometric cor-
ohmic losses) and the earth surface. In the case rections. The main objective in the scope of abso-
of the X-SAR system, this quantity is estimated lute radiometric calibration is the determination
before mission and measured in a receive-only mode of the overall SAR system transfer function rela-
during mission. On the raw data stage a noise ting image amplitudes Pi to a'-values, as expres-
subtraction takes place by using the resulting sed in the following equation:
noise data.
a' = 10.lg(P i ) - 10-lg(K') - 10.1g(rA rm) . (3)
6-7
40
30^
A
A
E
20
b A E-SAR (28.9±1.1) d8m'
b
0 .0 I I I I I I I a i
1 2 3 4 5 6 7 8 9 0 1 12 13
Number of Cornerreflector
Fig. 11 Amplitude accordance of 13 equally sized corner reflectors.
Pi is the image power in the case of distributed The2 mean values determined were 28.9 dBm2 and 29.0
targets and the impulse response minus an averaged dBm , which are in good mutual agreement and also
clutter (noise) level for point targets; rA deno- in good agreement with the theoretical value of
tes the azimuth resolution. K' is slightly diffe- 29.3 dBm2 .
rent to the factor K of equation (1) because of
the inclusion of the processing gain which is ach- Fully polarimetric SAR sensors, such as the DC-8
ieved by forming the
wit synthetic
refeenc taget aperture. hae
knwn CS-alue External
t be Ful
SAR anda thereSIR-C system,
SAs enr, suchoas require
additionally the t thee
reference
used targets
in order to determine
with knowntheRCS-values
calibration
havefactor
to be apiaino oaiercclbainpoeu
application of polarimetric calibration procedu-
K' nCorner
reflectors and active radar calibrators res. Such procedures imply the elimination of the
. Csensor induced distortions cross-talk and channel
(ARC) are appropriate tools to fulfil this task imbalance. This can be accomplished by utilizing
for the central and upper parts of the system either different types of corner reflectors or
transfer function,
ment whereas fin esecilly
situ RCS measure-
wih sctteometrs eleted eiter
clutterdi ticstogether
statistics of rer trihedral
with rel c corner
or
ments with scatterometers of especially selected reflectors [10).
area-extended targets are helpful to determine the
lower part of the system transfer function. S. GEOCODING
As an example, results of the 1989 Oberpfaffenh- Due to the sidelooking illumination geometry and
ofen airborne SAR calibration campaign are presen- the "azimuth-range-imaging" of SAR, mountains seem
ted in Fig. 9 and 10. These figures show the
C-band W overall system transfer functions of to bow toward the sensor. This phenomenon is known
DLR's experimental SAR (E-SAR) and NASA/JPL's DC-8 as foreshortening and, in the worst case, as lay-
SAR. Both system transfer functions represent best over. The latter happens in cases when mountain
fits of the corner reflector responses as marked peaks are less distant from the sensor than the
by triangles in the two figures. The calibration valley in front of them.
factors
an for better
accuracy both systems
than 1could be determined
dB, showing a fairlywith In Fig. 12 the foreshortening can easily be seen
good calibration performance for airborne SARs.The by comparing the projected slope area in front of
oo correspondence due to the SAR systems involved the mountain Sf, which appears in the SAR image as
has been determined to be within the 1 d error (R2 - Rl), and the projected slope area behind the
hasbee deermnedto
e wthi th I B eror mountain Sb, which appears in the image as (R3-
bounds for two typical earth surface object clas- mUn
ses: grass land and concrete. R2 ).
Cross-calibration becomes more and more important The power density relationship Sr between (R2 -
in order to keep the effort and expenditure on R,) and,(R 3 - R2 ) can be approximated as follows
multichannel and multisensor calibration within
tolerable limits. This is especially true in view Sf R3 -R2 (R1 +R2 )(R3 -R2 ) (D2-bi)
of the upcoming spaceborne missions, such as R2 -R1 b (R3 +R2)(R2 -R1 )(63 - 2 )
X-SAR/SIR-C and the follow-on activities. A measu-
re of cross-calibration quality is, for instance, The second and third factors in the numerator are
the RCS accordance of equally sized corner reflec- larger than in the denominator (see Fig. 12) and,
tors derived from image data from two different therefore, the power density ratio is >> 1.
SAR systems. Such an example of thirteen 0.9 m
corner reflector responses concerning the C-band This can be seen in SAR images in the form of
VV mode of the E-SAR and DC-8 SAR is illustrated bright slopes in front of mountains.
in Fig. 11 together with the theoretical RCS va-
lue. Geocoding means precise rectification. For this
6-8
;T!,
~
Fig.~~~ ~~ ~ 13GooedMtehonsee
6-9
SAR SIMDUATION
by
D. Hounam
Deutsche Forschungsanstalt fdr Luft- und Raumfahrt e.V.
Institut fUr Hochfrequenztechnik
8031 Oberpfaffenhofen
Germany
.....
......
7-2
2400
N.Rank: 9
1400
0 5 10 15 '20 25 30 is 40 45 so 50
*[eq]
.0 143.0 285.5 427.3 571.3 712 1 16 2 067.9114t'185,0
G o u d ran 9e [ ki ]
Fig. 1 Timing for the ERS-l satellite derived by the Cross-track Geometry Program. The wanted echo
window is identified by the vertical lines.
* Now the width of the received Doppler spectrum The bands are plotted on a graph of PRF versus
BD can be calculated: off-nadir angle (angle between the target and
the vertical) and ground distance (distance be-
2V e5 tween the target and the ground track). The axes
BD = X , (2) can also be swapped around. For the wanted swath,
the user can select the PRF, rank (number of
where V is the velocity of the sensor platform. transmit pulses after which the echo is received)
and the maximum and minimum off-nadir angles, i.e.
a To satisfy Nyquist, the PRF needs to exceed the the limits of the antenna elevation pattern. The
bandwidth: program calculates the position of the unwanted
ambiguous echoes enabling the requirements on the
PR > 1.2 B, (3 elevation pattern to be derived.
The above steps are simple approaches which can 3.1.2 Link Budget
require more sophistication, depending on the case Once the timing has been defined, the geometry is
in hand. In effect, a software utility needs to be known and the link budget can be computed accor-
developed to handle all aspects. For example, the ding to the radar equation. This is a straight
last step, where adequate sampling is considered, forward task and doesn't require much software
ensures that ambiguous responses in azimuth, due
to aliasing unwanted energy into the wanted Dopp- sophistication.
ler spectral band, can also be influenced by the The following form of the radar equation can easi-
characteristics of the azimuth compression routine ly be realised with a spreadsheet. Table 1 gives
in the SAR processor. However, the design of many an example with the values for the ERS-l satelli-
SARs has started down this path. te.
The timing of the received echoes requires consi- The radar equation for a single look in the final
deration of the cross-track geometry and the pulse
characteristics of the radar. The art is to find image can be written:
space between successive transmit pulses for the 2 -r X 2
=
echo with enough margin for the receiver to settle S/NLg. LNoJEc,
CP
[C±tcpPRFJ]
(4M
"
back to full performance. Another constraint is 0
that the echo from immediately beneath the sensor [A] [B] [C] [D]
(nadir) should also not fall close to the echo - r o r 1
window. Having an incidence angle of 900 it can be Llsani2-
strong enough to overload the receiver. Particu- [E] [F] [G] [H]
larly, for satellite geometries, many pulses may where,
be underway before the echo is received (for ERS-I
9) and just as many nadir echoes exist. These am-
biguous echoes interfere with one another and have G = transmit power, g
to be suppressed by the antenna elevation pattern. N. = receiver noise density,
An example of a software utility for analysing the TP = transmit pulse length,
timing is the Cross-track Geometry Program develo- tcp = compressed pulse length,
ped by the DLR. Fig. 1 shows a typical example for Tj = azimuth integration time,
the .tS-1 satellite. The program computes the va- PRF - pulse repetition frequency,
lid bands where the echo is corrupted by neither X = wavelength,
the preceeding or succeeding transmit pulses, nor r = slant range,
the nadir echo. Forbidden areas are shown hatched. Co - velocity of light,
7-3
j
0.00
0llll llill
A"1#1]6 no
M2s
. A A.F.... 6.4
0 ., *0o2 .o4i.1 69 Ol
-A 00l
ilpell.3
.Pp ob.
l I
0.l 24I A0
'ZIi-20 I
liep221..
l
P... 44
,Id 25 0
I 464
11:11 II1'$
-54
-1 p2
1 ii f .l 4Ioi 300
+3iii
Fig.llIimus 3ll~reposeTw-imnioa
The next phase of the Performance Estimator is to MUMSN 21.7 21.7 21.7 21.7 21.' 21.' 21.7
2.7 21.7
calculate the imaging performance for the various TOTAL 21.4 21.' 21.' 21.' 21.'231.7 21.7 31.' 1.7
PRF solutions. Normally, the three highest ranked
PRF solutions will be analysed but the operator
can change the ranking or the number of solutions,
if he desires. Table 4 Example of the Performance Estimator ambi-
guity performance.
on three screens:
Imaging performance is displayed
signal to noise, ambiguity ratios and target de- The target detection characteristics describe the
tection. ability of a SAR sensor to detect point targets
above a distributed target background.
Using built in algorithms for the radar equation
digitisation noise, bit error noise and a variety Man-made targets generally consist of single or
of scenarios for the normalised backscatter coef- combinations of discrete scattering centres and
ficient a., the Estimator computes the signal-to- can be regarded as having deterministic proper-
noise ratios for each selected PRF solution and ties. Such scatterers, often called point targets,
nine positions across the swath. The S/N at the are usually much smaller in extent than the re-
imput to the analogue-to-digital converter, i.e. solution cell size. Although the radiometric re-
the video SIN, the S/N for each processed look and solution can be calculated for such targets, the
the multilook SIN are all calculated. From the user is more interested in the ability to detect
latter the radiometric resolution can be computed. targets above the system noise or distributed tar-
The dynamic range is the range of signal powers get background (clutter).
over which a particular requirement, e.g. radio-
metric resolution is met. The Estimator searches To detect a scattering centre or point target suc-
over a range of a0 values until the radiometric cessfully it must be brighter than the brightest
resolution goes out of specification. Table 3 point in the speckled background.
shows typical results for the ERS-l satellite.
9 M1'02. an "N .doIn principle, the radiometric resolution can be
299
'S
4909
.
M0o4
1r-5
2 1useful
M calculated for point
parameter targets
if the radar and thissection
cross can beofa the
vm 2 o,
0 I1. :. 0: -0.9 -2.0 -3.4 target needs to be measured with known precision.
25. 20. 2. 0 24 0 29.9 23. 25.2 24.9 24.3
-1.1 41.2 41'. 41.. 41.1 ,.5 :o. ., .3 The application considered here, to image man-made
um 109 .92 3 .
1<'
.
.9 . -
s.' -0.3
1. 1. .30 4
-1.3
..
-2.9
-9.3
objects consisting of one or several dominant
L ,w .0 -1. scattering centres, requires that the weakest
1LON 1
I3 l. 19 1.3 0.9 .. -1 .3 -3.79
-1.o -.
L0W 4
59l
-3"1
t,
1.3 1.!
-1.2 -'1 1.1
1.8 -11 1J 5
-3.3 -04
-2,*'-,.1 -. 3 scatterer can still be resolved above the back-
I 0
2,.3 . 2.,3.3 .5 .4 ground. If sufficient scatterers from the target
rim"
.9 04 9.0 9.4 5.5 2. 9.0 4.3
3.7
4.o 3.'
7.9 9.0
1,0 14.3 -
-9 -19.0 -15.3 -151. -4.o-13.9 -1.0 can be detected, the size and shape of the object
W8 2. 21. 2 . 38
20. . 21. 28. 20.3 20.0 can be reconstructed and, hence, the object des-
RWXugW ,I:5 ,9 99 . 19i
Oo5 310M - 2,,947 cribed.
Table 3 Example of the Performance Estimator si- In the approach described here, two parameters
gnal-to-noise performance (ERS-l). from radar detection theory are used as a measure
of the ability to detect a scatterer. The false
The ambiguity performance in range is calculated alarm rate (FAR) is the probability that the in-
from the geometry, timing and antenna elevation tensity of the background in an image pixel is
pattern. Range ambiguities occur when the slant such that it Is falsly interpreted as a point tar-
range differs by half the pulse repetition inter- get. The detection probability is the probability
val (PRI). The ambiguity ratio is computed by in- that a point target can be detected above the
tegrating the energy in the wanted region and am- background.
biguous regions and dividing. The antenna eleva-
tion pattern can be entered aa a look-up table or A scatterer is considered to be detected in an
as a weighted pattern with up to seven weighting image pixel if the intensity of that pixel exceeds
coefficients, the intensity of the surrounding pixels by an
- *... .+ . . . . . .
7-6
imihk.d,
IM
smaller than the resolution cell, only the
point target contributes to the energy in the pkeibfmee FAt
corresponding pixel. This would apply to the w PD Ms-
case where the point target is surrounded by a bo
a surface of low scattering cross section on.na
large enough to fill the resolution cell and
where the system noise is negligible compared
with the background oo. This scenario, is ty- m
pical for large metallic structures, such as cows
aircraft, where the scatterers are surrounded T =oaowf% om ,t
by large metallic surfaces reflecting the en-
ergy away from the sensor. F OwaU f FAR
a variety of receiver gain setting algorithms the system analysis performed to date,
which use either centre swath signal, mean
swath signal and mean or the minimum/maximum * a listing of all system files available and
backscatter coefficient, as well as user se- the switching to other radar systems using
lected, function keys.
a choice of either fine or coarse range res- The Performance Estimator is typical of an algo-
olution, rithm based simulation tool. The software uses DEC
windows and allows the user to play with different
* various system hardware gains and losses, parameter settings and check the performance.
Clearly, the algorithms that are used are tried
0 the possibility of overlapping azimuth looks, and tested, so that few major surprises are to be
expected in the results. Full-blown simulators, as
a characterisation of the thermal and Analogue described below, mimic all the steps in the SAR
to Digital Converter (ADC) noise sources, system and can, hence, be used for scientific in-
vestigations.
0 a variety of amplitude weighting functions,
as well as user selected, 4. SIMULATION
0 the ability to degrade the range ambiguity A simulator attempts to model as many parts of the
ratio due to an expected range of radar back- SAR system without recourse to analytic functions.
scatter coefficients. The advantage of this approach is that genuine da-
ta are generated, which can be examined at each
The software characterising these functions has stage. Fig. 7 shows the elements of a SAR system
the following features: which many be simulated and the evaluations which
can be performed.
partitions which can be displayed on the
screen with descriptions, Although all elements contribute to the imaging
performance, some may be able to be simplified de-
DEC windows designed to allow the user to go pending on the application. For instance, if the
forwards and backwards through all the opti- application is concerned with the investigation of
ons to make and assess system changes quick- SAR data processors, a very primitive raw data
ly, generator with simple targets and ideal sensor may
be adequate for the task. For this reason simula-
the ability to generate antenna patterns from tors tend to grow, starting off being tailored to
a Fortran file, a particular application and later being augmented
to become a more comprehensive tool.
0 a facility to examine graphically the manner
in which, for example, the antenna patterns The first step in simulating a SAR system is to
and the ADC noise are modelled, model the target. The target can most easily be
simulated, if it is described by a two-dimensional
0 fast evaluation of a large number of swath matrix of scatterers each modelled as a complex
points which are split up into their azimuth vector voltage reflection coefficient. Fig. 8
look components, shows such a target model consisting of xy cells.
As will be seen, the simulator only has to form
0 all the timing information that can be paged the sum of all vectors within the instantaneous
for all the PRF and includes for example the field of view IFOV of the radar to generate the
predicted interference cases, time domain signal. The latter process is referred
to as raw-data generation and the final result is
0 tables and graphs of ambiguity, signal to to generate the signal received at the output of
noise and imaging data for all the selected the antenna. The raw data generation can be the
PRFs that are easily copied to a specified most complex part of the simulation chain, even
printer, exceeding the complexity of the SAR processor.
0 tables and graphs of required target size for Generating such a two-dimensional array of vectors
detection or detection probability versus assumes that the scattering process lends itself
false alarm rate for verification modelling, to this representation. Distributed targets with
differing normalised reflector coefficients ao
graphs of the components of the video signal can be synthesised with such a model, as can iso-
to noise ratio and the factors that degrade lated point targets. Mixing distributed and point
this, targets is also possible. The speckle associated
with distributed targets can be generated with an
a the analysis of the effect of changing any appropriate statistical model. Most simulators
system parameter and plotting the change pro- working in this way include utilities to help the
duced in any of the performance measures, user generate target patterns.
run information and a series of warning mes- The following description of the DLR simulator
sages when either algorithms are not functio- SARSIM will be used to illustrate such a simula-
ning or when the user has made a mistake, tor.
0 graphics software that allows the user to ea- 4.1 The System Simulator, SARSIM
sily select graphs, automatically scales and
grids the data, plots requirement curves and, SARSIM was originally developed by Ferranti in the
if applicable, fully labels the output and UK in 1977 taking about 15 man-years of effort.
provides a rescale facility for comparisons
with other graphs. In addition radar sensiti- The original software was extensively modified by
vity data are given under each modelled swath Ferranti in 1978/79, by Logica (UK) in 1981/82 and
point, currently by the DLR since 1978. Two versions of
the program exist, one on the DLR CRAY and more
book-keeping functions for the performance recently on a VAX4000 workstation. SARSIM is des-
plots which allow the user to keep track of cribed in [4 and 5].
7-8
Propagation Path
Attenuation
Polansation
Bd-errorsLink
Down
Target
Geometry
Composition
Scattering mechanism
EVALUATION
Image Processing
S Venfication of system design
Influence of system parameters
Investigation of SAR imaging properties
Development of interpretation algorithms
mV+ Vj = ko
where k is a constant. The target size is restric-
. .ted to 192,000 elements.
subcells valuea of
and the into
cell is divided the reflection
number coeffi-
(at least one) of
cient for each sub-cell is derived from the value
for the whole cell. This second stage permits the
..... introduction of speckle into the model, by assu-
,----, ming that the amplitude of each sub-cell value is
I ."
drawn from a Rayleigh distribution. The mean of
the Rayleigh distribution is specified by the cell
value generated in the first stage algorithm. The
,__________-- ____ "phase of each sub-cell value can either be fixed
or drawn from a uniform distribution (-x, x) Each
Fsub-cell value can be thought of as representing
the reflectivity from a point target located at
standard types of targets. Slope effects, such as the centre of the cell. Therefore, the target,
layover and shadowing, are not simulated. The ba- even though representing an extended target, con-
lyvp l sists of a matrix of point targets. This is a
sic principle of simulation is a "pulse by pulse standard method for modelling extended targets
simulation, that is the relative positions of c
platform and target, together with platform atti- Superimposed on this, the user can specify up
tude are calculated for each radar pulse and then to 16 deterministic point scatterers.
this is used to calculate the returned signal for Fig. 11 illustrates a possible target example. He-
each transmitted pulse. The returned signal is re- Fig ec illtas a pib tar exampl.
presented by a sequence of floating-point (com- re, each cell has been split into four sub-cells.
plex) numbers obtained by convolving the transmit- The amplitude, Ai of each sub-cell can be determi-
ted pulse form with the target, weighted by the ned in one of three ways:
antenna pattern and taking into account any spe-
cial receiver chain features. . Stochastic Background
The following
of the (first options are provided
stage) matrix for generation
of cells: The platform position and
Tepafrpsto n attitude are calculated
tiueaecluae
for every pulse in the time interval [-T/2, T/2],
A) User specified pattern: a two dimensional ar- that is the simulation is constructed time symme-
ray of amplitude data is read in from a file trically about time zero, time zero being the time
when beam and target are in alignment. In the case
provided by the user. of the elliptical orbit simulation, the orbit and
B) Cyclic pattern: a set of cycles is generated geometry parameters are sufficient to determine
in either the azimuth or range direction, the theoretical position of the platform for every
Each cycles consists of an alternating se- pulse. However, an additional perturbation may be
quence of high reflectivity and low reflecti- introduced in the
leration term. The form
flat ofearth
a constant linearrequires
acce-
vity amplitudes. The length of the sequences instead input data relating tosimulation,
the linear velocity
within each cycles is fixed, but increases by
one cell between successive cycle-series. A components at time zero and linear acceleration
example is given in Fig. 12. components before the platform position per pulse
can be calculated.
,st SNo 3The platform attitude per pulse is expressed in
CYCLE-SERIES CYCLE-SERIES CYCLE-SERIES yaw, pitch and roll angles. These can either be
AMPLITUDE -'f considered as constant or a.:!onforming to some
2 CELLS LS specified angular motion.
HIGH- I -C!
F)
flied amplitude at a specified position.
Checkerboard pattern: a pattern of alterna-
tic function of time.
4) a binary phase encoded pulse; the frequency
of the pulse is constant, but the phase can
I
ting squares/ rectangles of high amplitude take two values: 0 and x.
and low amplitude.
The phase structure of the pulse is specified by
G) Square of
square wellzero
pattern: a uniform
re~lectivtwy target with a
inset. binary string.
~jk -~ -
7-11
pulse, the
In all four forms of the transmitted
amplitude of the pulse is constant. -, ,
4.1.5 Antenna -, -
range gate
- m
The antenna is defined by a two-way antenna gain
pattern, specified in azimuth and elevation. .... i....
The antenna gain function is assumed separable, 1 . . m
that is the antenna gain in any given direction is
obtained by multiplying the appropriate azimuth 1 t T t. r 'targe
and eaNd elements
for gain variations during the simulation period.
The following gain patterns can be used in the si-
mulation: Fig. 13 Target element contributions to the mth
range bin return using signal generation
primary option (time domain).
1) Uniform pattern G = 1.
range bins. It should be noted here that the tar-
2) Modified cosine pattern, get cell range dimensions should be chosen equal
cos(x) wxd to, or smaller than, the range bin width. This is
G = 2I where x sine , to minimise problems occuring when few target
(S)2 - X2 cells happen to fall in a given range bin, a re-
sult of the fact that the target is in fact model-
where: D is the effective antenna aperture led by a number of discrete point scatterers,
(in azimuth or elevation), rather than by a true extended target.
X is the radar wavelength,
a is the azimuth/elevation offset The primary method of signal generation inherently
angle, includes all range migration effects due to earth
3) Sinc pattern, rotation, earth curvature and radar wave front
sin(x) curvature. The simulation of the range modulation
G (x) where x is defined above, can either be performed in the time domain (by
W 'convolving the returned signal directly with the
transmitted waveform), or in the frequency domain
4) User specified pattern. The user must specify (ymliligteaporaeFuirtasom
a table of gain against azimuth/elevation (by multiplying
and thenthetaking
appropriate Fourier transforms
offset angle.offst anle.together the inverse transform).
The beamwidths of the antenna pattern (in azimuth Two secondary options for received signal genera-
and elevation) must be specified by the user. The- tion exist, both of which attempt to reduce the
se values define the extent of the generated an- amount of computer time required. The first of
tenna pattern, these options generates the expected return (phase
history) from a point target located at the centre
4.1.6 Propagation of the target matrix assuming an ellipsoidal, lo-
cally spherical, earth's surface and an unpertur-
The radar equation is not included in the SARSIM bed elliptical orbit. Appropriate account is taken
model. Most of the terms in the radar equation are of the azimuth antenna pattern. The azimuth modu-
constants (for most system simulator applica- lation is performed in the frequency domain and so
tions), apart from the range gain dependence (r4 ). the Fourier transform of this phase history is
For satellite simulations the lack of range depen- used to calculate the return range gate values.
dent attenuation is thought to be of negligible The same phase history is used for all range gate
significance (though of more significance for air- values. Range migration effects are not accounted
borne applications). for in this model. Note that the target element
spacing is assumed to be defined by the range bin
Atmospheric distortions are only included in as width in range and the pulse repetition interval
much as the optional addition of a phase noise in azimuth.
term. The noise follows either a uniform, Ray-
leigh, exponential, or normal distribution; in the The second option extends the above idea to partly
case of the normal distribution an auto-correla- include range migration effects by taking into ac-
tion function can also be specified. The auto- count that the return from a point target appears
correlation function can either be linear, quadra- in several range gates, or alternatively expres-
tic, exponential, Gaussian, or user supplied. sed, that neighbouring target gates contribute to
the return from one range gate (Fig. 14). The
4.1.7 Signal Generation transition points between neighbouring range/tar-
get gates are calculated at the centre of the tar-
The primary option for received signal generation get (mid-swath) and assumed to apply to all ran-
generates the returned signal for each pulse in ge/target gates. The phase history segments are
turn in the time domain. The distance to each tar- convolved with the target in the time domain. In
get element is calculated and then the returns both of the secondary options the range modulation
(after weighting by the appropriate antenna gain) can either be performed in the frequency or time
from all target elements within a range gate are domain, as in the primary option signal genera-
summed, taking into account the appropriate pha- tion.
ses, to give the final return range gate value.
This operation is performed for all range gates 4.1.8 Receiver
within the sampled return pulse and for dll pul-
ses simulated and is, hence, not an insignifi- The receiver system typically consists of a low
cant task. Fig. 13 illustrates the target returns noise amplifier a down-converter, an IF amplifier,
which contribute to the return range gate value, and an IQ detector. The low noise amplifier is as-
sumed to be perfect apart from the addition of
The exact form of the curved radar range bin de- thermal noise. The down-converter is similarly as-
pends on the orbit and earth geometry simulated. sumed perfect apart from the possibility of a re-
The width of the bin is determined by the sampling sidual carrier frequency resulting from a mismatch
frequency of the ADC, which, together with the re- bet.een the up- and down-converters. This residual
ceiver gate times, also determines the number of frequency produces an additional term in the un-
7-12
range
' ' ' - ' '
gate
'
! ! '------"7-----
t- ----
I ' I . "-'-,targetI
elements , .. -t
]
muth directions, or on the transmitted chirp re-
Receiver system phase noise can be introduced im- plica.
mediately prior to the IF amplifier. The receiver
noise form (both phase noise and thermal noise The filter can either operate on both channels
from the low noise amplifier) follows either a (full quadrature) or just on one (real) channel
uniform, Rayleigh, exponential, or normal distri- (non-quadrature) by setting the remaining (imagi-
bution; in the case of the normal distribution an nary) channel to zero. The prefilter is assumed
auto-correlation function can be specified. The perfect apart from errors introduced by the finite
auto-correlation function can either be linear, word length within the digital filter. Word-length
quadratic, exponential, Gaussian, or user sup- effects (both before and during the filter) may be
plied, simulated by specifying the number of bits avai-
lable in the digital filter registers.
4.1.9 Analogue to Digital Conversion
4.1.11 Datalink
The Analogue to Digital Converter (ADC) can be
thought of as consisting of sampling and quanti- The datalink between the on-board system and
sing operations. The sampling operation is inhe- ground based processor can either take the form of
rent in the SARSIM model, as the analogue signal a telemetry link (satellite SAR) or a High Density
is represented by a set of discrete samples, Digital Tape (aircraft SAR). The model restricts
each sample consisting of a (complex) pair of the datalink simulation to allowing the user to
floating point numbers. The quantisation process specify a link transmission word length and then
then converts these floating-point numbers to a simulating bit errors over the link. The bit er-
finite set of quantisation levels consisting of rors can either be stochastic, where the spacing
both positive and negative values. Values excee- between bit errors is a random variable drawn from
ding the largest qunatisation level (ignoring an exponential distribution, or deterministic,
sign) are set to this level. The number of over- where a user-specified cyclic table (of up to 50
flows is counted. Quantisation of both the radar values) is used to determine the distance between
signal and transmitted chirp replica can be simu- one bit error and the next. The setting of the bit
lated. The quantiser itself can be specified in in error is switched. The datalink is simulated by
two forms: the following algorithm:
1. The user specifies up to 14 threshold values 1. Scale and quantise the data to the relevant
and corresponding output levels, link transmission word length, if necessary.
The scaling factor must be defined by the
2. The user specifies the number of bits n, the- user.
reby defining the number of quantisation le-
vp , (2n). The optimum equally spaced symme- 2. Insert bit errors.
i-ic quantiser is used as shown in Fig. 15.
3. Rescale the data to compensate for scaling
The quantisation process consists of the follov ng introduced in (1), if required.
steps:
4.1.12 Processing
1) Scale fhe data and add a DC offset, if requi-
red. The scaling can either be user-specified For a frequency modulated or "chirped" pulse, the
or automatic. For automatic scaling, the ma- first step in the SAR processing is to remove the
ximum, mean or mean squared, of the sar.ples frequency modulation. This is known as range pro-
in the first pulse, is adjusted (scaled) to cessing and can either be performed on-board in
be a user-specified fracticn of the maximum the analogue system using a direct replica of the
7-13
4.1.13 SARSIM Application Example series of small plates or facets enabling the tar-
get to be entered as a complex shape. The model is
In this section an :xample of a satellite SAR si- capable of handling hidden surfaces (radar sha-
mulation is given. The configuration of the system dows) and can generate polarimetric signatures, a
is based on that for the ESA Remote-Sensing satel- feature of increasing interest for the next gene-
lite (ERS-l). Errors and noise sources within the ration of SAR sensors.
system are not simulated in this example.
5. CONCLUSION
Fig. 17 shows the radar target used in the example
simulation: a single point target on a non-reflec- Parametric analysis and simulation software packa-
ting background. Fig. 18 is a plot of the range ges for investigating SAR systems have been des-
compressed SAR data. Fig. 19 shows the final image cribed and their merits discussed. The usefulness
plot of the single point target. of simulators is often called into question as the
cost and manpower invested can be large. It is,
therefore, worth considering two applications of
0.0, the SARSIM simulator.
50.0 # ERS-1
Imp 05.A0 N
S IO EOO-IO..i~. Dy.AW ft'U
0.0
A-ALDUI IN OARDMANG
G1191511"6. REFERENCES
f'
Z'
11597&008
RNLGEONBADPotter,
000000TSIDN
MANGE
2TINE 5.5DLR-IB
INI
31 Hounazo, D.
S.
Schmid, R.
Algorithms and Software
User Guide for the Perf or-
mance Estimation of Syn-
thetic Aperture Radars.
551-4/92.
1 'EE2
Proc. IEEE '86 National
--A- ALIRT"
:2: II
i~g.0I..1.0Riccio,
/
I I 81 Franceschetti, G.
Migliaccio, M.
D.
Prentice Hall 1975.
SARAS: A Synthetic Apertu-
re Radar (SAR) Raw Signal
Simulator.
Schirinzi, G. IEEE Transaction on Geo-
Fig. 23 2-bit quantisation noise curve, science and Remote Sensing
Vol. 30, No. 1, Jan. 1992.
For both the 2 bit case (Fig. 23) and the 5 bit
case (Fig. 24), the simulated values are much less
than the theoretical results. An at least partial
explanation of this is the non-white nature of
quantisation noise [11] and the influence of the
7-16
it
__ _ _ _ __ ___ _ _ i
8-1
Abstract
The Shuttle Imaging Radar-C (SIR-C) is the third in a series of space shuttle based synthetic
aperture radars (SAR) sponsored by the National Aeronautics and Space Administration
(NASA). The SIR-C ground data processing system is to process the playbacked SAR signal
data into a variety of data products for distribution to the science community. This paper
presents an overview of the end-to-end ground data processing system with emphasis on
the unique characteristics involved in the system design. Included in the discussion are
science requirements, radar system specifications, input data format specifications, system
operations design, data products design, processing algorithm design, hardware architecture
design and software design.
1. INTRODUCTION ucts. Table 2 shows the orbit characteristics and the radar
specifications.
will for the first time
The Shuttle Imaging Radar-C (SIR-C)
provide simultaneous data acquisition of eight radar chan- This paper presents the system design of the SIR-C ground
nels (two radar frequencies: L and C, each comprised of data processor, which is being developed at the Jet Propul-
two like-polarized and two cross-polarized channels) from a Table 2: SIR-C orbit characteristics and radar
spaceborne synthetic aperture radar (SAR) (Curlander, 91a], specifications.
[Curlander, 91b], [Jordan, 91]. This instrument is accompa-
nied by a VV-polarized, X-band SAR (X-SAR) which will SAR Orbit
operate simultaneously with SIR-C. The SIR-C/X-SAR is -Nominal Atitude 215 ± 25o Km
-Eccentricity s 0.002
currently scheduled for three space shuttle flights, the first - Inclination 57*
being in late 1993. Each flight is planned for a six to eight Attitude Measurement Error(3o)
day data acquisition period. Table 1 summarizes the SIR-C - Roll ±1.24"
science requirements on the image quality of output prod-
Yaw ±1.430
- Pitch ±1.73
Table 1: SIR-C science requirements on the image Attude Drift Rate Eror(3o)
quality of output data products. Roll +0.03-1soc
- Yaw ±0.03°lsec
- Pitch ±0.031saec
Reslution broadening s 20% Transmitter frequency
Integrated Sidelobe Ratio 1 -14 dB L-Band 1.25 GHz
- C-Sand 5.30 GHz
s -17 d8
Peak Sidelobe Rato Polarization HH, HV, VH. VV
II
8-2
sion Laboratory (JPL) under contract with NASA [Curlan- multi-band, polarimetric SAR image [Freeman, 89], [Klein,
der, 91b]. There are two European space agencies working on 92]. For radiometric calibration, the built-in-test-equipment
the X-SAR processor development: the German Aerospace (BITE) data are designed for probing the health of the an-
Research Establishment (DLR) and the Italian Space Agency tenna, receive only noise data for estimating the noise power,
(ASI) [Runge, 90]. calibration tone signal for monitoring the receiver gain and
temperature measurements and T/R module failure informa-
The major challenge to the SIR-C processor design is to cope tion transmitted via the downlink telemetry. These ancillary
with a large number of radar modes. Nominally the SIR-C data and calibration site data are essential to derive radio-
science team has selected seventeen data acquisition modes metric calibration parameters which are applied during the
from all the possible combinations with eight radar chan- data processing to produce calibrated image product. For
nels (See Table 3), two pulse bandwidths and three data geometric registration, special consideration is required in
quantization formats. Additionally, data will be collected in the processor design to ensure that the output images are
two nominal attitudes over incidence angles from 170 to 630 registered in both cross-track and along-track dimensions.
with a variety of antenna elevation patterns controllable to
provide beam spoiling at the steep incidence angles. The The remaining paper presents an overview of the SIR-C end-
large number of radar modes complicates the logics in han- to-end ground data processing system design, which includes
dling different types of data format and increases the scope input data format specifications, system operations design,
of testing the integrated processor software, data products design, processing algorithm design, hardware
architecture design and software design. At the end of the
The second challenge to the SIR-C processor design is to paper, we give a brief summary of the status anld plan for
attitude
cope with the large attitude uncertainties and high
drift rates of the space shuttle platform as shown in Table the processor development.
2. The large attitude uncertainties create PRF ambiguity
problem in Doppler centroid estimation and large Doppler er-
rors for data acquired over high terrain relief areas. Special 2. INPUT DATA FORMAT
techniques (multiple PRF technique and attitude steering
technique, respectively) are required to resolve the problems The signal data is recorded across four recorder channels on
[Chang, 92a), [Chang, 92b]. The high attitude drift rates in- the High Density Digital Cassette (HDDC). The data rate
duce fast Doppler drifts in both cross-track and along-track for each recorder channel is 45 Mbps for a total of 180 Mbps.
dimensions. Frequent Doppler update is required to main- The signal data is quantized into one of three types of for-
tain the image quality, which complicates the geometric rec- mat: 4-bit, 8-bit or (8, 4) block floating point quantization
tification procedure to produce a seamless image. (BFPQ). Nominally, the data is collected over a period called
data take using the same set of commanded radar parame-
ters. The length of the data take varies from 3 minutes to as
The third challenge to the SIR-C processor design is to pro-
duce radiometrically calibrated and geometrically registered long as 15 minutes for ocean site data. The average length
is estimated to be between
4 and 5 minutes.
I,
Table 3: SIR-C radar data aquisition modes. The nominal SIR-C data take consists of a turn-on sequence,
followed by the science data collection and a turn-off se-
WaE CHANNEL-1 CHANNEL-2 CHANNEL-3 CHANEL4 quence as shown in Figure 1. The first four seconds of the
I tHH IHH LHV IHV turn-on sequence consist of (one second each): receive only
2
3 LVH LVH LVV LVV
044 0l C4V CVV noise data, caltone scan data, low noise amplifier (LNA)
4 CVH CVH CVV CVV BITE data and high power amplifier (HPA) BITE data.
5 UH UHH LVV LVV These four second data are used for radiometric calibration.
6 044 01H CVV CVV The remainder of the turn-on sequence consists of one sec-
7 LHH LHH CVV CVV
a tHH w
C"1 044 ond of PRFA data and one second of PRFB data which
9 LVV LVV CVV CVV together with the first second of PRFc data are used for
10 1 LV LVH GVH I CVH resolving PRF ambiguity in estimating the Doppler centroid
II UH LM 044 COf
12 LVH LW CV4 CVV frequency [Chang, 92a]. The system remains on PRFc for
13 U4H LVV 04 CVV collection of the science data. The turn-off sequence
S '14 UNH LYN LHV LVV sqec is simi-
14 . LM 2HV WL. lar to the turn-on sequence in that the science data collection
Is LVH, LH LVV, UHV CV, CHN CW, CHVi is followed by one second each of PRFV and PRFA. The
T last four seconds of the turn-off sequence are receive only
T=~~ CUM, I CVV noise data.
8-3
Figure 1: SIR-C input data run format. Each segment in turn-on and turn-off sequence is 1 second duration.
RON: Receive Only Noise, LNA: Low Noise Amplifier, HPA: High Power Amplifier, CAL scan: Caltone Scan.
At every one second time tick, a null-line is inserted. The 4. DATA PRODUCTS DESIGN 4
null-line is obtained by setting a half of the phase array el-
ements with a 1800 phase difference to create a null around The SIR-C output data products include three image prod-
the center of the antenna elevation pattern. The null-line is ucts: survey image, standard multi-look image and standard
used to estimate the shuttle roll angle drift, single-look image; and one reformatted signal data product.
The throughput requirements are
to produce 24 survey im-
A sinusoid waveform, called the calibration tone (caltone) age products per week during the phase 1 operations and
signal, is injected in the receiver electronics and recorded to produce 9 standard multi-look image, 1 single-look image
together with the return echo data. The caltone is used to and 1 reformatted signal data products per week during the
estimate the receiver gain change as the temperature varies, phase 2 operations. The expected processor throughput far
exceeds the requirements.
3. SYSTEM OPERATIONS DESIGN
The survey image is a 4-look, single-polarization strip image,
Operations of the SIR-C processor is comprised of two main stored in the byte amplitude format. The image is deskewed
phases: phase 1 survey processing and phase 2 standard pro- to zero-Doppler and resampled to the ground range domain
cessing, which last for a total of one year. During the phase 1 with a 50 meter pixel spacing. The resolution is approxi-
operations, a quick-look survey processor is employed to pro- mately 100 meters. The length of the survey image is equal
cess single-frequency band, single-polarization channel data to the length of the data take. The average length is approx-
into low resolution strip images. These survey images will imately 4.5 minutes or 2000 Km. The survey image will be
cover all the SIR-C ground sites albiet with a single radar recorded on Alden thermal prints and CD-ROMs. The CD-
channel. By-products of the survey processor include un- ROMs will be distributed to all the principal investigators
ambiguous Doppler centroid estimates history and roll an- (PIs).
gle estimates history. Additionally, during the phase 1 op-
erations, some selected data segments (covering calibration The standard multi-look image is a multiple look, polarimet-
sites) will be processed into single-look, full-resolution com ric (single-, dual- or qnad-pnlari7Atinn) frame image. The
plex imagery. These data will be analyzed to derive the image is deskewed to zero-Doppler and resampled to the
parameters used for antenna pattern generation and polari- ground range domain with a 12.5 meter pixel spacing. The
metric calibration. These parameters will be applied during azimuth resolution is chosen to be 25 meters. The range res-
the phase 2 standard processing to produce phase and am- olution is chosen to be 25 meters or the natural resolution if
plitude calibrated data products [Freeman, 891, [Klein, 92]. greater than 25 meters. The image data is stored in a com-
pressed cross-product format [Dubois, 89]. The basic frame
The system operations schedule is planned as follows. Six size is chosen to be 100 Km. The image will be recorded on
weeks are allocated for processor check-out upon receipt of Kodak prints and CEOS formatted tapes.
the first signal data tape. Phase 1 operations will begin
following the completion of the processor system check-out The standard single-look complex image is a single-look, po-
and last for a period of twelve weeks. This is followed by larimetric (single-, dual- or quad-polarization) frame image.
phase 2 operations for a period of forty weeks. The image is processed to full-resolution, deskewed to zero-
I
s-S.
8-4
Doppler and presented in the slant range domain in natural is performed using the spectral analysis (SPECAN) algo-
pixel spacing. The image data is stored in a compressed rithm which requires fewer azimuth FFT's than the tradi-
scattering matrix format. The basic frame size is chosen to tional matched filtering algorithm. Following azimuth com-
be 50 Km. The image will be recorded on CEOS formatted pression, radiometric correction is applied to compensate for
tapes and a reduced, detected image will be printed by the the along-track radiometric modulation. This is followed by
Kodak printer, a geometric rectification step that resamples the slant range-
The reformatted signal data contains the signal data refor- Doppler image into the ground range cross-track and along-
matted in the range line byte format. The signal data to track domain. The rectified burst images are then overlaid
pouetefnlmlt-oksrpiae
to produce the final multi-look strip image.
gether with the decoded radar parameters will be stored on
CEOS formatted tapes. For the survey processor, the initial Doppler centroid fre-
quency is determined using a clutterlock routine and a ambi-
5. PROCESSING ALGORITHM DESIGN guity resolution technique that requires a multiple PRF data
collection at the start of each data take [Chang, 92a]. The
5.1 Survey Processing Algorithm unambiguous Doppler centroid frequency is then tracked by
a burst mode clutterlock algorithm during the data process-
The SIR-C survey processor utilizes a burst mode process- ing. The Doppler frequency rate is solely derived from the
ing algorithm [Sack, 85], [Curlander, 91b]. The algorithm ephemeris parameters. Analysis results show that the accu-
flow chart is shown in Figure 2. The survey processor is de- racy of the ephemeris is sufficient for generation of survey
signed to process an entire data take into a strip image in products without employing the autofocus routine.
approximately one-seventh the real time data collection rate.
To attain high throughput rate, the data is bursted in az-
imuth (slow time) with a one-quarter duty cycle factor. The
data volume is further reduced by a factor of four in range Prior to standard processing, preprocessing is employed to
(fast time) by processing the data using only one-quarter iiteratively refine the Doppler centroid frequency and the
of the range chirp bandwidth. The azimuth compression Doppler frequency rate estimates using clutterlock and aut-
ofocus techniques [Li, 851. Doppler centroid frequency is
Selecled Null Lines Bursted estimated from the azimuth spectrum by locating the en-
segment Adjacent Lines Range Lines
of data ergy centroid. Doppler frequency rate is estimated from the
look registration error by azimuth cross-correlating the look-
Caltoe Null-Line RneWak 1 and look-4 images obtained by spectral division. Identical
P Ing Poel Doppler parameters are used for processing all polarimetric
gain angle Range ange data channels to ensure the phase coherency required for the
estimate stimate Reference Comprsslon polarimetric data analysis. This approach will result in some
Cross-Track Generation
Radiometric oincrease in azimuth ambiguities if the antenna beams are not
i Turn exactly aligned.
Tlj
8-5
Migration|
6. HARDWARE ARCHITECTURE DESIGN
Single-look
Full-aperture
Azimuth I Reference
Iutniaid I Figure 4 shows the hardware architecture design of the SIR-C
Reference Iground data processor. The entire processor system is com-
Generation Azimuth posed of seven subsystems. The Data Transfer Subsystem
nvers FFT I (DTS) performs raw data reformatting and line synchroniza-
Single-Lool Multi-Lool with eight compute elements and an Alliant FX/2800 mini-
Complex Complex supercomputer with twelve i860-based CPU's. The STAR
Multi-Lool array processor is the main compute engine for survey proces-
Detected sor. Its FFT performance is measured at 120 MFLOPS using
Oversampling Oversampling three computational modules. The Alliant FX/8 is primar-
and Slant-to- and Slant-to-
Ground Range Ground Range ily used for standard poetprocessing functions. Its aggregate
Conversion Conversion FFT performance nears 20 MFLOPS. Two SKYBOLT accel-
SLo - erator boards are installed to speed up the FX/8 computer,
Reduction Filtering
Isatoi
(intensity
II Filtering
(cros'productl
which provide additional 100 MFLOPS compute power. The
Jdt )atrin r FX/2800 is the main compute engine for standard processor
matrix dat) data) data)
Tta
and standard preprocessor. Its aggregate FFT performance
is measured at over 300 MFLOPS. Computational tasks are
Reduction
(cross-product distributed over computers for concurrency processing in or-
data) der to provide maximum processor throughput.
Dat
D Transfer Subsystem (OTS)
__ IF24]I .6'
m Cutalog Subystm AS WN
LAWa iA 401_411
eam
"I
data quality and integrity. The OPS consists of Exabyte rically registered. Another challenge to development of this
tape drives, Alden thermal printers and Kodak color print- large software based system is its complex interfaces among
ers. Three subsystems, OPS, GAS and CAL, run on three the many software programs. Clear interface definitions are
separate SUN Sparc workstations. The image display and essential to successfully deliver the operational system on
~operator interface display are handled via X-ternminals. schedule.
1'
~uses both FORTRAN and INGRES.
8. SUMMARY
nmjosotrparwaeoftis
tion.pocesiTe opeatins nlte 993
he ignl
SACKNOWLEDGEMENTS
from the ancillary data in order to ensure that the output [Chang, 92a C.Y. Chang and .C.Curlander, "Applications
inmge products are radiometrically calibrated and geomet- of thlleMule PRF Technique trvete is Doppler Cetrod
8-7
Estimation Ambiguity for Spaceborne SAR, IEEE Trans- Transactions on Geoscience and Remote Sensing, Vol. GE-
actions on Geoscience and Remote Sensing, September 1992. 22, No. 6, November 1984, pp. 592-597.
[Chang, 92b] C.Y. Chang and J.C. Curlander, "Attitude [Jordan, 91] R. Jordan, B. Huneycutt, M. Werner, "SIR-
Steering for Space Shuttle Based Synthetic Aperture Radar, CX-SAR Synthetic Aperture Radar Systems, Proceeding
Proceedings of 1992 International Geoscience and Remote of the IEEE, Vol 79, No. 6, pp. 827-838, June 1991.
Sensing Symposium, Houston, May 1992. [Klein, 92] J.D. Klein, "Calibration of Complex Polarimet-
ricSAR Imagery Using Backscatter Correlations, " IEEE
[Curlander, 91a] J.C. Curlander and R.N. McDonough, Syn- Transactions on Aerospace and Electronic System, Vol. 28,
thetic Aperture Radar: Systems and Signal Processing, John No. 1, pp. 183-194, January 1992.
Wiley and Sons, 1991. WLi,85] F. Li, D. Held, J.C. Curlander, and C. Wu, "Doppler
[Curlander, 91b] .1C. Curlander and C.Y. Chang, "Tech- Parameter Estimation for Spaceborne Synthetic Aperture
niques in Processing Multi-Frequency Multi-Pol. Spaceborne Radars, " IEEE Trans. on Geoscience and Remote Sensing,
SAR Data, ' European Transactions on Telecommunica- Vol. GE-23, No. 1, January 1985, pp. 47-56.
tions, Vol. 2, No. 6, pp. 605-617, November 1991. [Runge, 90 ] H. Runge and R. Bamler, "X-SAR Precision
[Dubois, 89] P.C. Dubois, et al, "Data Volume Reduction for Processing, " Proceedings of the International Geoscience
Imaging Radar Polarimetry, " IEEE International Sympo- and Remote Sensing Symposium, College Park, Maryland,
sium on Antennas and Propagation, 1989, Vol III, pp.1354- May 1990.
1357. (Sack, 85] M. Sack, M.R. Ito, and I.G. Cumming, "Appli-
[Fe n 8cation of Efficient Linear FM Matched Filtering Algorithms
[Freeman, 89] A. Freeman and J.C. Curlander, "Radiometric to Synthetic Aperture Radar Processing, " lEE Proceedings,
Correction and Calibration of SAR Images, " Photogram- Vol. 132, Pt. F, No. 1, February 1985, pp. 45-57.
metric Engineering and Remote Sensing, Vol. 55, No. 9, W
September 1989, pp. 1293-1301. [Wu, 82] C. Wu, K.Y. Liu, and M. Jin, "Modeling and a
Correlation Algorithm for Spaceborne SAR Signals, " IEEE
[Jin, 84] M. Jin and C. Wu, "A SAR Correlation Algo- Trans. on Aerospace and Electronic Systems, Vol. AES-18,
rithm which Accommodates Large Range Migration," IEEE No. 5, September 1982, pp. 563-575.
I:
isi
1'1
I 1
9-1
J.P. Hardange
Thomson-CSF
178, Bd Gabriel P6ri
92242 MALAKOFF Cedex
FRANCE
= ,
... ~ ...~~~....
. ....... ..
....
.... . . . .. ... ...... . ....
..
.......
2..
9-2
The filter, which is matched to a signal of The first derivative of the ambiguity
Doppler parameters fa, fa, fa, has the function at (0,0) is equal to zero, because
following impulse response: at this point, the function is reaching its
t
2
.. t
3 maximum:
h(t)-s*(-t) exp [-j2n(f. t + f. -- + f. -- I dIX(0,f)1 2 I
2 6
(24) df I
if-o
- 0 (33)
Notations:
The second derivative can be computed, as a
f "f, - f. (25) function of the value of the ambiguity
function at (0,0):
- fo - f. (26) d2Ix(O,f)j1 I_
-2(2n) 2
- - - - - - -- ----- f t Is(t)['dt[x(0,0)j
df' Ifo (34)
*- (27) f0 is the half width of the peak on
the
Output of the matched filter: Doppler axis, at 3 dB:
2
Ix(0,f0 )1 1
X( ,f,f,f f. (T-t)s,(t)dt (28) ---O -
2
2(35)
IX(0,0)1 2
X( ,f,f,f ,f.,f.) - s(t) s*(t-r) eJ# dt The width of the peak is then:
* I I..(29)
with: 2 f .....-- Is(t)- dt)/2 -- (36)
t 2 t
3 21 (ftls(t)2 dt)1/2 T.
- 2 it fo t + fo -- + fD -- (30) T. is the "equivalent duration" of the
2 6 signal. For the signals of finite duration,
which are considered in our application, T.
(t- )2 . (t-r) 3 is very near from the duration of the
- f. (t--) - f signal.
2 6
-_----.-
-
I~Eape
9-4
duin
6' 0,78
an aprue h f
x~ it4ftepaki
t4 Is(tfl 2 dt
2 I(t)1
16
=
2
hn
X(0,0,0)1
dt)/ 2
2
(42)
(3
2
R T. T. itui dt)1/
Example:
3.2. Resolution in Doppler acceleration For an observation during an aperture time
We make now the hypothesis that the Doppler Ta, the resolution of equation (43) gives:
frequency is null, and we use the same 85 5.7
principle to evaluate the resolution in 2 0 - ------ (4 i
f~ IT, T2
IX(f)P
,O - (T ~ 2 jift 2
4.1. Description
(38) The radar is fixed (point A of figure 1).
The target is rotating at a constant rate
w. The center of rotation is C. The range
dIX(0,0,f)121 between the radar and the center of
-Ix(0,0,0)P1+f ------------V. rotation is R.
dt jf-0 We consider one reflecting point of the
target (point M of figure 1), which is
located at coordinates (x,y,0), at initial
f2 d2 jX(0,0,f)jIj time (t=0). x and y are small compared to
-+ ---------- 1. (39) R.
IX(0,0,0)1 2 - --
2
f
if dt2
"t 4 IS(t)
If-0
2
dt (40)
The coordinates of M at each time are
(u,v,O): o t- in t(5
y,
V
T U
M (X,
, 0)
1< AC X
R y sin w t target.
- RC-- -
--- (50
(50 ) The image is a projection in a plan
R°
0 containing the radar-to-target axis (range
The Doppler frequency of N is then: axis) and the cross-range axis which is
orthogonal to the axis of rotation.
2
fn- - w (y cos o)t + x sin w t) (51) Figure 2 gives several examples of basic
projections.
For a short observation time T. around t-0, - In figure 2a, the target is oriented to
the Doppler frequency is depending only on the radar and has a pitch motion. The
the y coordinate: range resolution is on the lenght axis of
the target, and the cross-range
2 yu resolution is the height of the target.
- (52) The projection plan is a range-height
X vertical plan.
We have established a direct relation In figure 2b, the target is oriented
between the Doppler frequency and the perpendicular to the radar range axis and
position of the reflecting point on a has a roll motion. The range resolution
cross-range axis, relative to the center of is on the width axis of the target, and
rotation. the cross-range resolution is still the
height. The projection plan is a
We can note that the reception filter has width-height vertical plan.
to be matched to signals of constant
frequency and finite duration. As a In figure 2c, the target is oriented to
consequence, the ISAR processing is reduced the radar and has a yaw motion. The range
to a simple Fourier transform, resolution is on the length axis of the
target, and the cross-range resolution is
the width. This case is comparable to
4.2. Resolution SAR. The projection plane is a
length-width horizontal plan.
As a result of equation (36), the
resolution in Doppler frequency, for an
observation time T., is:
In several applications, the configuration
k is not so simple. The target principal axis
rfD - -- (53) are any possible orientation compared to
T. the range axis. There can be simultaneouly
pitch, roll and yaw motions. The
where the value of k is near to 1. orientation of the instantaneous axis of
This result is well-known for every rotation is not known. Consequently, the
conventional spectral analysis. It could be position of the projection plan is not
known either. I
resolution, by
possible to achieve a better
the mean of estimation methods.
Unfortunatly, these methods require a high
signal-to-noise ratio and a large amount of 5- RELATIONS WITH SAR, REAL APERTURE, AND
computation. They are not currently used in TOMOGRAPHY
this field of application.
5.1. Relation with sAR
The resolution on the cross-range axis is
then: Figure 3 shows a typical SAR configuration.
The radar is flying at constant speed, u,
X rfo X and height on a straight line. It is
ry - - (54) observing a first reflecting point designed
2 w 2 w Ta by N The azimuth of the target is #1. The
Doppler frequency is:
r . (55) 2u
2 ** fD .... cos # (57)
with
There is a second reflecting point, 32.
- c Ta (56) which is at the same range, but with a
*, is the angle of rotation of the target slightly different azimuth:
during the observation time (figure l).It 2 * + 64 (58)
is interesting to note that the resolution
.9-6
a Height
Pitch C~p
Range= LMht
R~oll N D =Height J
Flange Wdh
=Wkdth
Dopper
Yaw AxQ
Range= L-ht
2u
f,2 Cos (+J + 64) (59) __
x U
2u
f'1 - fD2 . ...-sin *, 64 (60)
2u 1
--- sin #, r# - -- (61)
.T.
r# - (62)
2 u T. sin +,
The projection of this angular resolution
at the range of the target is:
S-- (63)
r 2 u T3 sin #1 (
L - u T. (64)
9-7
L u Ta
(65) 1 - 2 L (70)
R R
The reouinr ca
The resolution
function of :r can be written as a 5.3. Relations with tomography
- Rotation angle refered to the radar: SAR Let a(x,y) be the function to map.
- Rotation angle refered to the target: The measurement are made by the mean
ISAR of several 1-D projections on
different directions. The direction
We have established this result, of the line-of-sight for one
considering a short observation time. We projection is #. The value of the
have
order limited the computations to first
developments: projection at each range u is:
p*tu) - alx(u,v,#),y(u,v,+)] dv
- In SAR technics, it is the hypothesis of . (71)
the unfocused SAR. The processing is the
same (Fourier transform), whatever the with notations of figure 1:
range of the target is: it is focused at
infinite range. x(u,v,0) - u cos + - v sin + (72)
- In ISAR technics, the equivalent name y(u,v,#) - u sin 4 + v cos 4 (73)
"unfocused ISAR" is not usual. But the
principle is the same: second and higher The projection function has a
order terms in the variation of the range Fourier transform:
* are neglected.
For a longer apeture time, there is a P#(U) - p(u) e du (74)
difference between SAR and ISAR:
The Fourier transform of the
- In a pure ISAR configuration, the range projection function is one slice of
between the radar and the center of the 2-D Fourier transform of the
rotation of the target is constant. image a(x,y) (projection-slice
theorem). The orientation of the
- In a SAR configuration, this range is slice is given by the angle *:
varying following a quadratic law.
However, the effect of this range P#(U) - Z(U cos #,U sin #) (75)
ij variation can be compensated through an
appropriate demodulation of the received Z(X,Y) is linked to a(x,y) by the
signal. relation:
E(MY) a (x,y) • dx dy
ft I
9-9
0 0'YZ
R
j_
tI
9-10
fo A
azimuth
shift
t
1/ fr
Roll motion period
W " 3*/s
6.2. Characteristics of the image
If we realize the filter described by T, - 1,5 s
equation (31), the resolutions in azimuth
and height can be estimated using equations - 450
(37) and (44), (88) and (89). Resolution in azimuth: r. - 5,2 m
The best result will be obtained around
t-0. At this time, the Doppler frequency in ResolutiOn in height: r5 - 1 m
is null).(88)Considering
equation to .y (velocity
is due onlyU>>U in
(the term of Taking into account all the approximtions
l othat have been done to come to this result,
the aircraft greater than the velocity of these values are to be considered only as
the ship) the resolution in azimuth is: orders of magnitude.
X f. (114)
y - / 2 ..
w (109)
(109) /.
e *(u)-#(v)] -j2xf(v-u)
2IX(0f) - e du dv
2 x xVariable change:
Ix(o,f)l X...... (11o)
ix0f)I(1)w - v-u (116)
4
IX(o,0 1
IX (O,0) I
IX (0,0) I =
IX (0, fm) I 2 xo
- fM fm Doppler Frequency
f e ti(u-w+ufdu -
fr I+j[#(u1-#(w+ul]
WAEFR
GENERATION
UP-CONVERTER
2
I [+(u-#(w+ul
- du
2 (118)LO
T. - T. #2 + R++(w)
(119) RECEIVER
2 2
Ix(0,f)1 - T6(f) -T.# 6(f) + S (f)
1to) [JDOWN CNERTER
4x 1
S/L X 2 ------
P()2 (123) 8- REFERENCES (unclassified publications
BA, T. onlyl
E,(f) - 4 sin 2 (n f Tt) £(f) (125) 6. Walker J.L., "Range Doppler imaging of
Rotating objects", IEEE Trans.
- 2 /c is the time corresponding Aerospace and Electronics Systems,
to the range of the target. VOL. AES-16, Nol, Jan. 1980, pp.
23-53.
The effect of this weighting is an
enhancement of the high frequency
noise power, by a factor of 2, and 7. Dike G., Wallenberg R., "Inverse SAR
an attenuation of the low frequency and its Application to Aircraft
noise power, by a factor: Classification", IEEE 1980
International Radar Conference Record,
e(f) - 4 V f2 T,' (126) Arlington, April 1980, pp. 161-167.
II
9-14
ji
I
10-1
The requirements, capabilities, and limi- Signals encountered in SAR analysis have
tations of single beam moving target phase as their most important attribute, and
indicztion (MTI) for a SAR are reviewed, thus must be coherent over each sequence
Special processing for multi-look SARs to of received samples. Coherence in the
enhance SAR ocean wave imagery contrast azimuth dimension may be exploited to
and directional spectral estimation is achieve remarkably good resolution, and
highlighted. may be used for special techniques as well.
The classic treatment of SAR analysis is
Since the SAR signal is coherent, signal that of Harger [1], who makes the usual
phase comparison is possible between pairs assumption for most of his treatment that
of signals. For spatial separation of the the two SAR dimensions, range and azi-
signal pair, interferometric signal combina- muth, are uncoupled, and that the antenna
©CanadianCrown Copyright
L - -
10-2
I
10-3
image is complex, having a phase structure platform, these are sufficient to establish
that may be exploited for certain applica- the Doppler bandwidth of the signal. The
tions. Ideally, the complex image is simply class of sensors includes SAR, certain
a linear transformation of the input signal, SONAR devices, Spotlight SAR, tomogra-
and has no phase errors introduced by oper- phy, inverse SAR as used for imaging Earth
ations in the processor. (Note that very few satellite and nearby planets, etc. The
SAR processors in use today satisfy this azimuth resolution is given by the inverse
objective.) Selected advantages of the (spatial) bandwidth, which leads to
complex image domain for calibration _.
am.-
10-4
SAR Application: For a SAR, the subten- in which there is no dependence on range
ded angles of view are generally very small, or wavelength! This is a fundamental
and, furthermore, are oriented nominally at characteristic of SAR systems, and makes
right angles to the velocity vector of the them well suited to spacecraft platforms.
sensor. In this case the small angle sine NB: The magnitude of expression of Eq. 12
approximation applies, and one obtains is an approximation only, as it violates the
X correct diffraction limited result that follows
PSARZ 2 P,
- (9) from Eq. of
benefits 7,and it does
orbital not account
geometry. It for the
is this
I
10-6
Antenna size for a Spotlight SAR is gov- In the early 1970's the "stretch" technique
erned by the desired size of illuminated was demonstrated [15]. For a linear fre-
area, and by the required signal-to-noise quency modulated (linear fin) signal, de-
ratio of the received signal sequence. Both modulation of the received signal by a
of these objectives are range dependant, delayed replica of the original results in a
leading to larger antenna size for larger difference signal of constant frequency.
ranges, all else equal. Phased array anten- The frequency is proportional to the
na technology is highly desirable for such relative delay of the received and reference
systems, although impressive performance signals. Useful results occur only when the
may be obtained with rather modest scan signals overlap substantially.
angles. For example, from Eq. 7, one may
find that resolutions on the order of only 5 Stretch is perfect for the Spotlight appli-
times the wavelength (15 cm at X-band) cation. It reduces the bandwidth require-
may be achieved with an illumination angle ments needed for all portions of the system
of about 3° (in the nominally side-looking following demodulation. It does this by a
Spotlight geometry). For an N-look image, trade-off between range and bandwidth, a
I
J2
10-7
convenient consequence of the linear fm The solid lines in the figure represent the
waveform. It is thus restricted to scenes (two-way) power profile of stationary scat-
with relatively small range extent, which fits terer return as limited by the azimuth pat-
the Spotlight requirement. Stretch modula- tern of the radar antenna. The zero order
tion applied to the Spotlight SAR case is Doppler response is shown together with
described by Walker in [13]. two of the ambiguous spectra each centered
on a multiple of the PRF. The width of the
3.4. MTI clutter spectrum is B. = 2PVIX, where p is
the two-way half-power angular width of the
An ability for moving target indication antenna pattern, and V is the radar vehicle
(MTI) has been for many years an objective speed.
of both strategic and tactical radar systems.
It is of interest to explore the ability of a The dotted lines represent the envelope
SAR in this application. Early work in this returned from a single target having radial
area may be found in [1], and especially velocity v with respect to the radar line of
[161, both of which are restricted to rates sight, leading to a Doppler frequency shift
rather small comparedto that of the radar 2v/A which appears as an additional phase
platform. For larger motions, the problem term (normalized to azimuth spatial fre-
rapidly gets more complicated [17]. quency by V1) in Eq. 4. (This may be
derived from from Eq. 1 by using
Attempting to simultaneously achieve both R. = > R.+ vt.)
fine spatial resolution and spectral reso-
lution flies in the face of physical principals We are interested in describing the way a
(e.g. [181). The general formulation of this slowly moving target appears in a SAR
principle is through the ambiguity function. image, and in the possibility of detection of
These fundamental limits apply to the con- moving targets.
ventional SAR configuration, suggested in
Figure 2. For a SAR, the small (coherent) radial
velocity component of a scatterer leads to a
shift & in the mapped azimuth position of
IT its image. This is a natural consequence of
the fact that the azimuth coordinate system
is derived from Doppler information, and
*,* the processor has no way of knowing th.,
the target itself might have a Doppler com-
ponent. For the aircraft case [16], this
argument leads to Ax = R, v/V. Since the
• .1 range to a scatterer is relatively large, there
is often a shift in scatterer position many
0
PRF 2 times the azimuth
less there resolution.
are tell-tale signs However,
of where un-
the
Figure 2. SAR Doppler space in the object should be located (such as the wake
presence of a moving target. (Two of a ship), it is impossible to identify the
ambiguities explicitly shown.) return as being associated with a moving
target. This is an example of the spatial-
Doppler ambiguity inherent in the problem.
10-8
I
promise has yet to be fulfilled. tional spectrum derived from the SAR data
[25]. Furthermore, the technique naturally
_I
,J "
10-9
The underlying assumption of two pass of slope continuity through which progres-
interferometry is that the terrain being sive phase shift may be integrated across
observed is essentially unchanged (with changes in excess of 21r. Use of low
respect to phase characteristics) in the time resolution methods such as "shape from
interval between observations. The shading" is also helpful.
condition may be inverted, since the
absence of interference fringes is an Within the last decade, interferometry for
indication of (randomly) changed phase in SAR is one of the two most significant
the affected area. developments. (Quadrature polarimetry is
the other one.) The topic remains an area
The basic SAR interferometric principle of very active research, and may reasonably
leads to terrain height estimation with a be expected to offer valuable quantitative
precision on the order of the resolution of results to the user community in the years
the SAR [28]. Its accuracy, however, is very to come.
sensitive to the value of the baseline dh.
For applications in which a known "level" 4.2. Temporal Interferometry
terrain feature is present, such as a
shoreline, then the elevation contours may There is an alternative way in which to
be calibrated. In principle, the technique build an interferometer. For a SAR on a
may be used with any satellite SAR. moving platform, a pair of antennas could
be arranged to lie along the flight vector,
The measurement capability of interfero- thus providing essentially identical views of
metric SARs has been extended through the the illuminated field, but at slightly
use of differential techniques [29], [30], [31]. different times. The basic geometry is
Through this approach, one looks for illustrated in Figure 4.
differences in the fringe pattern between
two pairs of observation opportunities. In
effect, in each case, one of the signal pair
establishes a reference phase for each
element in the scene, and the other signal
provides an estimate of phase change with [
respect to the reference. In principle, the md
method is sensitive to physical changes in D
position of a reflecting element to less than
)/4, which has been demonstrated [29]. It l
has been proposed as a method for estima-
ting crustal movement either before or after
an earthquake [301. Figure 4. Dual antenna time-sequential
p. I
..
10-11
The time delay 6t between the pair of where, neglecting noise, one may show that
received signals is DIV if the antennas are
operated separately as transmitter and MA ( )
receiver. The time delay is D/2V if, as is 1 5
usually the case, only one antenna is used (15)
for transmission and both are used for
reception.
as the indication of the presence of a
Scene elements having radial motion v,ad moving target. The detected difference
cause a differential phase shift image will contain only moving targets (plus
differential clutter and noise), each
A0 - 4 n vra 8t (13) weighted by the squared magnitude of the
X radially dependent phase term of Eq. 15.
If ssl(t) is the signal from the front antenna, + mm(t) + exp( -j4v -a 8t)]
and ss 2(t) is the signal from the rear
antenna, and mm(t) is the signal from a (16)
slowly moving target with radial velocity
* -a..-
-.- .-
10-12
The two components of the moving target Improvement of image quality, both through
are shared proportionately between the two summation of extra non-coherent looks, and
through coherent integration in special
channels, with sin2 (kv,.5t) in the circumstances, has been demonstrated using
separate data sets, available, for example,
difference channel, and cos2 (kv,.6t) in from nearly spatially coincident orbits.
the sum channel, where k = 2w/A. These By taking the two-pass topographic inter-
expressions show that there is a "blind ferometric geometry to the limit, in princi-
speed" vr,, B,at which the target motion is ple one could generate multiple range offset
through one (round-trip) wavelength data sets sufficient to create full holo-
between observations, and thus not graphic (three dimensional) images of the
detectable by phase shift. Blind speeds are illuminated surface. In fact, this was
characteristic of MTI radar systems. demonstrated nearly 20 years ago using
aircraft based radars, but has certain
Energy partitioning of the moving target physical limitations when presented as an
signal between the sum and the difference
channel may be used for target velocity optical analog.
modern It could
era of digital have value
imaging, in the
however.
estimation, analogous to the same technique
used for calibration of dual channel Both spatial- and time-delay interferometry
polarimetric radars [36]. have been implemented using a SAR. The
i
10-13
Under the constraint that phase of the [7] C. Cafforio, C. Prati, and F. Rocca, Full
signal ensemble is available, then there are Resolution Focusing of Seasat SAR Images
many varied measurement possibilities open in the Frequency-Wavenumber Domain,
to the clever inventor. One may reasonably IntemationalJournalof Remote Sensing, Vol.
expect that in the forthcoming years, more 12, 1991, pp. 491-510.
special SAR techniques will be developed,
and be adopted as needed for specific [8] R. Stolt, Migration by Fourier Trans-
applications of this exciting device, form Techniques, Geophysics, Vol. 43, 1978,
pp. 49-76.
1 -
1I-14
[15] W. J. Caputi, Jr, Stretch: A Time- [23] R. K. Raney and P. W. Vachon, Syn-
Transformation Technique, IEEE Transac- thetic Aperture Radar Imaging of Ocean
tions on Aerospace and Electronic Systems, Waves from an Airborne Platform: Focus
Vol. AES-7, No.2, Mar 1971, pp. 269-278. and Tracking Issues, Journalof Geophysical
Research, Vol. 93, No. C10, 1988, pp.
[16] R. K. Raney, Synthetic Aperture Radar 12,475-12,486.
and Moving Targets, IEEE Transactioinson
Aerospace andElectronic Systems. Vol. AES- [24] P. W. Vachon and J. C. West, Spectral
7, No. 3, May 1971, pp. 499-50f. Estimation Techniques for Multilook SAR
Images of Ocean Waves, IEEE Transactions
[171 E. J. Kelly and R. P. Wish"-r, Matched on Geoscience and Remote Sensing, to
Filter Theory for Hign-velocity Accelerating appear, 1992.
Targets, IEEE Transactions on Military
Electronics,Vol. MIL-9, Jan 1965, pp. 56-59. [25] P. W. Vachon and R. K. Raney, Ocean
Waves and Optimal SAR Processing: Don't
[18] J. R. Klauder, The Design of Radars Adjust the Focus!, Transactions of the IEEE
Having Both High Range Resolution and Geoscience and Remote Sensing Society, to
High Velocity Resolution, Bell System appear, 1992.
Technical Journal,July 1960, pp. 745-808.
[26] L. C. Graham, Synthetic Interfero-
[19] R. K. Raney and R. A. Shuchman, meters for Topographic Mapping, Proceed-
SAR Mechanisms for Imaging Ocean ings of the IEEE, Vol. 62, No. 6, 1974, pp.
Waves, Proceedings 5th Canadian Sympo- 763-768.
sium on Remote Sensing, Victoria, B.C.,
1978, pp. 495-505. [27] H. A. Zebker and R. M. Goldstein,
Topographic Mapping from Interferometric
[20] C. T. Swift and L. R. Wilson, Synthetic Synthetic Aperture Radar Observations,
Aperture Radar Imaging of Ocean Waves, Journal of Geophysical Research, Vol. 91,
IEEE Transactions on Antennas and Propa- No. B5, 1986, pp. 4993-4999.
gation, Vol. 27, No. 6, 1979, pp. 725-729.
[28] C. Prati, F. Rocca, and A. M.
[21] W. R. Alpers and C. L. Rufenach, The Guarnieri, Effects of Speckle and Additive
Effect of Orbital Motions on Synthetic Noise on the Altimetric Resolution of a
Aperture Radar Imagery of Ocean Waves, Interferometric SAR (ISAR) Surveys,
IEEE Transactions on Antennas and Propa- Proceedings of the International Geoscience
gation, Vol. 27, No. 5, 1979, pp. 685-690. andRemote Sensing Symposium, Vancouver,
Canada, July 1989, pp. 2469-2472.
Raney, W. J.
[221 K. Hasselmann, R. K.
Plant, W. Alpers, R. A. Shuchman, D. R. [29] A. L. Gray and P. Farris-Manning,
Lyzenga, C. L. Rufenach, and M. J. Tucker, Two-Pass Interferometry with Airborne
Theory of Synthetic Aperture Radar Ocean Synthetic Aperture Radar, IEEE
Imaging: A MARSEN View, Journal of Transactions on Geoscience and Remote
Geophysical Research, Vol. 90, No. C3, May Sensing, to appear, 1992.
1985, pp. 4659-4686.
I
10-15
R. Keith Raney
Canada Centre for Remote Sensing
system. This is understandible, particu- to a massive short circuit in the slip ring
larly when technical change as well as assembly of the solar panel primary power
language and terminology confound the system. (There are system evaluation
issue. In this review, primary sources are reports available from the committees
used whenever available. Furthermore, charged with investigating this event,
when there is an important conflict in should any reader be interested in
published data, comments are included in hardware and programmatic details.)
the text to help clarify the matter for the Data provided by Seasat proved to be of
interested reader. high quality and immense interest to the
science and applications communities, and
still appears as prime material for recent
3. PREVIOUS ORBITAL SAR SYSTEMS papers in the professional literature.
For some individuals, history hinges on the In the Table, the antenna is described as
wisdom of Lao Tzu (1st century, China), "corporate". This is a shorthand notation
and for others on the inspiration of Jesus for a passive antenna using power dividing
Christ (1st century, Judea). For those techniques to distribute the signal to be
interested in synthetic aperture radar, transmitted to a network of radiating
however, history hinges on Seasat (20th elements. In the case of Seasat, the
century, North America). Virtually all antenna was a flat microstrip array, built
Earth observing SAR satellite systems in on eight panels. The size of the antenna
this lecture owe both their inspiration and for Seasat is of interest. Its long length,
baseline performance standards to Seasat, and large aspect ratio, is "typical" of
and by implication its design and science "standard" (satellite) SAR antennas. The
teams. Of course, there are solid logical reasons for this may be found in [3], and
reasons for the technical similarities, but in other lectures in this series.
the fact remains, Seasat led the way.
Seasat was initiated by an inter-agency Typical parameter values important for
study team, and its implementation was image quality for satellite SAR systems, in
managed by the Jet Propulsion Laboratory particular resolution and number of looks,
(JPL) (of the California Institute of were firmly established by Seasat. Unless
Technology in Pasadena, California) one is willing to give up other image
funded by NASA. Seasat was a magnifi- aspects such as swath width, there is
cent achievement [1]. rather little flexibility available.
Principle parameters of Seasat are The NASA objective for Seasat was digital
sketched in Table 1, abstracted from [3], processing, but NASA ran out of money
also available from a widely distributed (after the potential processor contractor
literature. As in the remainder of this had been selected, but before the contract
lecture, most comments in the discusssion had been awarded!). The project was
on this SAR are keyed to entries in launched, literally, with optical processing
pertinent columns of tables. as the baseline. In parallel, Charles Wu at
JPL (and MDA, Canada) developed digi-
The design lifetime of Seasat was two tal processing techniques for Seasat.
years. Unfortunately, the spacecraft failed Seasat was a resounding success largely
after three months of SAR operation due because of digital processing, as the
I
1~1
11-3
remainder of the Tables will support. After the conclusion of the Kosmos 1870
mission, Glavkosmos started to advertize
Seasat used an analog downlink with a its SAR products on the international
bandwidth of 20 MHz. The data was market, using agreements with such
recorded, thence a/d converted (5 bits, I organizations as Space Commerce Corp-
and Q, fixed point) to an effective data oration (USA). Kosmos 1870 was a radar
rate of 110 MB/sec. limited in several regards by its "analog"
foundations. Image products from that
The Shuttle SAR missions extended the radar appeared to be far inferior to
foundations established by Seasat in the expectations based on paper specifications.
dimension of incidence angle. Both SIR- Optical processing in series with analog
A and SIR-B [2] provided valuable results, data links seem to have significantly
but with technical limitations. SIR-A used reduced the effective bandwidth of the
only optical processing, depending on data, leading to fewer looks and more
airborne film recorders adapted to the coarse resolution than the radar itself
Shuttle mission. (SIR-A was an should have provided. (N.B. "Digital"
achievement in its own right, however, as image products available from Kosmos
it was the first payload to fly aboard a 1870 are digitized optically processed SAR
space Shuttle platform.) For SIR-B, two data, not to be confused with digitally
problems interfered with the SAR data processed SAR data.)
quality, including a partial failure of the
signal line connector to the antenna (with
severe increase in noise level), and failure 4. PRESENT SAR SATELLITES
of the data downlink system (TDRS
antenna tracking system) so that the data In contrast to the almost total absence of
recovery originally planned was curtailed. (civilian) space radar capability in the ten
Never-the-less, much valuable data was years following Seasat, the present decade
collected within the constraints of the is witnessing substantial activity. All
flight mission. known space based SARs presently in
service are listed in Table 2.
For two years at the close of the decade,
the USSR operated the Kosmos 1870 4.1. Lacrosse
radar [4] [5], at the time a classified
system. Kosmos 1870 was one of three Although a classified mission, there have
satellite SARs built in the late 1970's built been several open literature publications
by NPO Machinostroenye. (The other two (e.g. [6] and [71) from which the entries in
are known at present by the name of the table for Lacrosse have been gleaned.
Almaz, discussed below.) The Soviets had The key items include resolution, down to
a series of well known real aperture radars one metre [7], and the antenna size, which
in space (the Kosmos 1500 series), but may be estimated from an illustration
Kosmos 1870, in addition to Venera 16 published in [6] and Paris Match (date
and 17 to Venus, was its first publicized unknown). These two facts lead to the
orbital SAR. Kosmos 1870 and its Almaz conclusion that Lacrosse uses the Spotlight
relatives are the only space radars with SAR technique in its high resolution
two antenna assemblies, directed to mode. This requires a steerable antenna
opposite sides of the orbital plane. pattern, so it follows that the rather large
I
TABLE 2. Present SAR Satellites
I
11-6
antenna must be a phased array. Orbital the heart of which is a SAR whose
parameters are based on the capability of parameters are listed in the table. After
the Shuttle Atlantis launched from Cape several months in the three day repeat
Kennedy, for which 570 is the limiting orbit used for system verification and data
inclination. (Lacrosse is manoeuverable validation, the orbit was changed. The
[7], however, so that current values of first two weeks of April 1992 used the "roll
inclination and altitude may differ from tilt" mode in a 35 day repeat pattern. The
those in the Table.) The press reports at "roll tilt" accomplishes a change in
least two other Lacrosse systems awaiting incidence angle to about 35'. After mid-
operations, one of which may already have April, ERS-1 resumed normal operations,
been launched. Imagery from Lacrosse is still in the 35 day repeat orbit. The 176
not available. day repeat orbit will be used for an
extended period later in the mission to
4.2. Almaz support altimetric experiments. Data is
downlinked when within range of a
Almaz (11, [41, [5]), or "diamond in the receiving station on X-band at 105
rough", isvery similar to Kosmos 1870, but MB/sec using 5 bit quantization (fixed
up-graded in several regards. Data is point) on the I and Q channels. Image
stored onboard in four parallel video quality from ERS-1 is excellent.
recorders whose capacity limit the length
of each data take. The data downlink is With two operational SARs in orbit, it is
analog for both realtime data and interesting to compare their respective
recorded data. The most significant views of the Earth. Figure 1 shows an
improvements in the Almaz signal chain area near Whitecourt, Alberta as seen by
followed from conversion to digital data both Almaz and ERS-1. The contrast
handling and processing. The standard between these two images is striking.
processor is patterned after a VAX 780, ERS-1 appears to be more sensitive to
and uses the conventional range/Doppler topographic effects, and less sensitive to
algorithm. Image quality is variable, but differences in vegetation. Almaz, on the
some excellent examples are available, other hand, easily displays the clear-cuts in
the forest, but is less sensitive to the
From an applications point of view, Almaz terrain relief.
is a very interesting system. Its S-band
wavelength and selectable incidence angle 4.4. J-ERS-1
make it a good choice for certain Earth
resource issues. An example of Almaz The Japanese launched J-ERS-1 ("Earth
imagery may be found in Figure 1. Resources Satellite") early in 1992. After
early difficulty with the mechanism, the
4.3. ERS-I SAR antenna was successfully deployed on
9 April. At this writing, the SAR system
After extensive studies and preparations is still undergoing inflight checkout; the
from about 1975, the European Space initial imagery is excellent. J-ERS-1 is a
Agency launched ERS-1 in 1991. The joint development of the Ministry of
primary payload [8], [9] onboard ERS-1 International Trade (MITI) and the
("European Remote Sensing" satellite) is Science and Technology Agency of the
the Active Microwave Instrument (AMI), National Space Development Agency
11-7
,
p
oI
i0
I
-- .. ...
..
--
r11-9
The radar bands selected represent orbital from this radar system. These, coupled
SAR heritage (L), current operational with a degree of freedom in bandwidth
preference (C), and extension to a new and incidence angle selection, lead to very
(civilian) orbital wavelength (X). The complex planning for the operations of
antennas required have been matched in SIR-C/X-SAR. In addition, there are
two regards. First, in order to allow the options in the number of bits to be
same PRF for all three bands, dictated by included in the signal data path, and
the requirement to have simultaneous allowance for experimental passes in the
multi-frequency operation, each of the ScanSAR format (see RADARSAT
antennas should have the same length. below), and for squinted beam data
Second, in order to have nominally the collection. (At this time, however, there is
same elevation beamwidth, the height of no plan to support data processing for
the antennas must be scaled in proportion these special imaging geometries.)
to their wavelength. The size of the total
array is about 12m by 4.1m. The three Processing for SIR-C, and particularly for
antennas will be mounted on one common X-SAR, poses its own challenges. Normal
structure in the cargo bay of the Shuttle, Shuttle angular motions, in combination
vertically stacked. with Earth rotation, lead to relatively large
and time varying Doppler centroid
Antenna technology for the X-band system variations in the received SAR signals.
is slotted waveguide, horizontally For the Shuttle, pitch, roll, and yaw are
polarized. Elevation steering is by allowed to vary within 1.5'. (The attitude
physical rotation of the antenna radiating control system for the Shuttle is not
surface about its longitudinal axis. "proportional", but depends on angular
momentum impulses from gas-jet thrusters
Antennas for both the C- and L-band to correct angular position when the
radars use active phased array technology, design "dead-zone" is exceeded.)
There are 18 panels for each antenna, Tolerance by the coherent radar signal to
a each one consisting of transmit/receive
(T/R) modules for H and V polarizations,
(yaw) attitude is determined primarily by
azimuth antenna beamwidth. For a given
By using phase control of the individual aperture size, the beamwidth of the
T/R modules, the antenna patterns can be antenna pattern decreases in proportion to
steered in both elevation and azimuth. wavelength. For example, the azimuth
beamwidth for X-SAR is about 0.150,
The H and V phased arrays, backed up by which is ten times smaller than the Shuttle
parallel receiver and data recording attitude uncertainty. It follows that
chains, may be cycled to achieve reception Doppler centroid estimation and tracking,
of the complex scattering matrix of the and its compensation in time and range, is
scene, the so-called quadrature polar- the most challenging SAR processing task
ization technique (described in another among all others represented by the
lecture in this series). SIR-C/X-SAR will mission. This circumstance has led DLR
be the first time that this capability is to new frontiers in SAR data processing
available from space. techniques, Le., [12] and [13].
A
11-13
Standard 25 x 28 4 100 20 - 49
Wide (1) 48-30 x 28 4 165 20-31
Wide (2) 32-25 x 28 4 150 31 - 39
Fine resolution 11-9 x 9 1 45 37 - 48
ScanSAR (N) 50 x 50 2-4 305 20 - 40
ScanSAR (W) 100 x 100 4-8 ° 510 20 - 49
Extended (H) 22-19 x 28 4 75 50 - 60
Extended (L) 63-28 x 28 4 170 10 - 23
Beams
(nominally) 45' . A narrow swath results to image a section of one sub-swath must
from the requirement to minimize be from consecutive pulses in order to
beamwidth in order to maintain good provide adequate sampling, and must be
signal to noise ratio, and also from the of sufficient length to allow formation of
necessity to maintain data rates consistent the synthetic aperture needed for the sub-
with downlink channel capacity. Wide swath at the required resolution. The
swath modes are supported by wider imaging operations are therefore split up
antenna beam widths than normal at into a series of blocks of pulses, each
steeper incidence angles, and use of the block providing returns from one of the
smallest available range pulse bandwidth sub-swaths. Each block is processed to
leading to more coarse ground range provide an image of a section of the
resolution. Signal to noise ratio and data corresponding sub-swath. The imaging
bandwidth arguments apply in these operations cycle around the full set of sub-
modes that are counterparts to those for swaths sufficiently rapidly that the imaged
the fine resolution modes, but with the sections in any one sub-swath are
result of broadening the usable swath adjoining or over-lapping.
width. Extended modes result from
selection of beams outside of the nominal RADARSAT will be the first operational
500 km accessibility region, either closer satellite radar system to implement the
to nadir (steeper incidence), or further ScanSAR technique.
away (more shallow or grazing incidence
angle).
6. AIRBORNE SAR SYSTEMS
5.2.2. ScanSAR
This section provides a glimpse at civilian
In order to allow imaging of a swath much airborne SAR systems now in service.
wider than ambiguity limits would Since most of these radars are meant for
normally allow, the RADARSAT system technology development as well as
has been designed to incorporate an applications experiments, the hardware is
alternative and less conventional mode frequently changed. The parameters listed
([15] and [16]) known as ScanSAR. In in Table 5 are thought to be an accurate
this mode, for which rapid steering of the representation of the basic performance of
elevation beam pattern of the antenna is those systems. Additional comments, and
essential, extended range coverage can be a resume of other airborne SARs are
obtained using a set of contiguous beams, included in the following sub-sections.
enabling images of total swath width up to
about 500 km to be produced. This is 6.1. Comparison of Selected Systems
accomplished at no increase in mean data
rate from the radar, but at the cost of The most widely deployed airborne SAR,
degraded resolution of the resulting image. and the one having the most modes, is
that of NASA which carries the radar of
The principle of ScanSAR is to share JPL (the Jet Propulsion Laboratory). This
radar operational time between two or radar, known as AirSAR ([2] and [19]), is
more separate sub-swaths in such a way as relatively new, having been designed and
to obtain full image coverage over their built to replace its predecessor which was
combined swath. The set of returns used destroyed in an aircraft fire in 1985. The
11-15
Ii
11-16
quadrature polarimetric capability at the do the previous two systems. Its recent
three frequencies of this radar offers a extension to X-band is in support of
unique and very rich data source that is DLR's vested interest in X-band for the
made available by NASA to investigators X-SAR radar. The system includes a
around the world. AirSAR also offers two quick-look (QL) onboard processor having
other special modes that merit attention, 50 m x 50 m resolution.
as noted in the Table. InSAR is an inter-
ferometric mode created by using data The radar of the Technical University of
from two antennas, one mounted above Denmark (TUD) has been designed to
the other on the side of the aircraft. The offer a variety of incidence and image
interference pattern between them may be parameter values within the constraint of
used to deduce terrain height information, being a single polarization C-band system
For AirSAR, this capability is at C-band [24], [251. Within the limits set by their
only. The other special mode also respective imaging geometries, the radar is
requires two separate antennas, this time matched to the ERS-1 SAR, aided by the
spaced along the line of flight, use of an aircraft capable of high speed
Interferometric measurements with these and high altitude.
two antennas may be used to observe
phenomena in the scene that change over The leading civilian SAR that is used for
the short interval At between observations, mapping surveys is STAR-i, owned and
such as currents on the ocean's surface. operated by Intera, of Calgary, Alberta
[261. This X-band system has performed
The SAR flown by CCRS (the Canada more than 75% of all of the airborne
Centre for Remote Sensing) is on a radar mapping done for commercial
Convair-580. Both X- and C-bands are clients world-wide since 1986. Data
fully supported by onboard real time products from this system are digitally
digital processors [20], and have a variety rectified to map accuracy standards, and,
of modes and data combinations available, with the recent use of GPS, allows
The standard image products are pro- accurate mapping with no need for
duced at seven looks. The system has surveyed control points. Under satellite
recently been modified to incorporate full navigation control, the radar map itself is
quadrature polarimetry on C-band, and the most accurate source.
may be operated in an InSAR mode.
Signal data is recorded so that ground 6.1. Overview of Other Systems
processing may be used for specific
experimental purposes, such as inves- There are several other airborne SARs
tigations requiring access to separate looks that deserve mention. The Netherlands
for optimized oceanographic SAR wave for many years has been supporting the
imagery. development of an advanced system
known as PHARS [27]. Currently,
The airborne radar of DLR (defined in PHARS is a C-band SAR mounted in a
section 5.1) continues to be upgraded with Swaeringen Metor II, a twin engined
new modes and capabilities [211, [22], and business plane. Early results meet or
[231. It is designed primarily for high exceed specifications. Nominal resolution
resolution and technology development, is about 5 metres with about 6 looks. The
hence it has a narrower swath width than antenna is W polarized to support
11-17
experiments with ERS-1. The program is II has already been built, and is virtually a
committed to implementing a full twin to Almaz. There is an intention to
quadrature polarimetric version within a launch and operate this radar, but the
few years. changing infra-structure of the former
USSR space segment places the future of
One of the pioneers in the field of SAR is Almaz II in doubt. Progammatic responsi-
the Environmental Research Institute of bility has been assumed by Russia for
Michigan (ERIM), whose civilian PRIRODA ("nature"), a complex payload
experimental SAR is managed by the that includes a SAR as one of several
United States Navy Air Development instruments [17], [18]. Much of the
Center (NADC), and is mounted on a P-3 hardware for the PRIRODA payload has
aircraft. The radar operates at X-, C-, and already been built.
L-bands, and is fully polarimetric [28].
Nominal resolution is in the 3 metre The EOS-SAR (Earth Observing System
range. SAR) [21 is at a different stage. There is
an active design and science team at work
The French have been actively involved in at JPL on EOS-SAR, but the program has
SAR for many years. Their SAR (a yet to receive funding approval from
civilian version of a former Thomson-CSF NASA. The Proposal by the EOS-SAR
classified X-band system) is known as team to NASA is to be in 1993.
VARAN-S [29], and has supported several
remote sensing experiments in Europe. The EOS-SAR as presently conceived
Polarizations are HH and VV. Nominal would be a most ambitious system. It
resolution is on the order of 5 meters, would carry the multi-frequency,
with four looks, nominal, quadrature polarization capability of SIR-
C/X-SAR into a long design life satellite
In addition to these relative well known implementation.
systems, others might be mentioned for
completeness. The Chinese Academy of In addition to these SARs, there are other
Sciences developed their own X-band initiatives being developed. France for
SAR several years agb. The system is several years has been promoting SAR-
called CASSAR, and has seen limited de- 2000, an X-band space radar intended to
ployment in the Far East. The USSR complement the SPOT program. Recent-
maintained an airborne SAR capability as ly, France and Canada have entered a
a test-bed for satellite programs, both bilateral agreement to do joint studies of
Earth oriented and planetary. India has a combined satellite radar system building
developed its own airborne SAR as well on the SAR-2000 and the RADARSAT III
as an indigineous digital SAR data planning. RADARSAT II, essentially a
processing capability, replacement for RADARSAT, is under
consideration by the Canadian govern-
ment. The European Space Agency has
7. PROPOSED SAR SATELLITES sponsored several studies of SAR concepts
that look beyond ERS-2. Undoubtedly,
There are three SAR satellites likely to be other space agencies have exploratory
approved for completion and launch in the space radar studies under way.
next decade, as noted in Table 6. Almaz
I
1I1-18
ii
11-19
I
11-20
BIBLIOGRAPHY
St
I
B-2
4k 640
C 4 -- V lo
0. V6 0C '.> 0 ±, 41IC1
0~~~> 0.'4 . 6 ooo o
0.34 V 06 0 V *60k00 .4
Ak S 3"a 41664,
.0 0 .4 -4 4j -4 *-K CO .0.02 IO0 0 4
0.-
0
4.
~ w
44 a0-4
k40 054
6 410 k.-41
1..
- IV NCI r 6k C0026,
-46Eoc m o V6
4.woo .6- r
4) tCrc W4 -- 0 4 -460 g-IC *
.0~~ kW a. a. "046 V k 0 4
V 1" C4 A-
41<64- D4% 0 - % -6.606 0 2k
%-- x
O6CV4 0:wI a 6,02
r..36~ v02X0.$ 0 66666
04 0 .2 -0 6026kg Z . I0 %
.- 41 U kOA02s
v4 - 41-4 4
0 0 6) 0 OkH 6 0 ;6 N
C 064-
4402 a-60 4A 1 41 -4 464..100" 0 z 0. 1464
NU40 4W8 '-0 X021 IDZ).
4 0. 604to0 600
N4
' -0 -0;W - 00 6 0 .
41 0- C4.s.2-0
02 .0 X "-40
60.64>0A o -4D a 66C k&4 600
dF0
m
Q0.
U4
2 I
j 14
160
.0 A441-
0100
4,
1
o
40
a4
cl
Pr0
wk 0- 6- 02 r 6...46
4 0 -- 6 C
444 0. -. ".~ 1- -.
0 .4 o F4 1., 0.041 Z 0.. 004 0 J4J
cikO
V0 a2 9 , .A A 04" .
,t 14 0 00 ar 5 V
U)C4
R-3
0.4j 0-S
M>.-
-.0 6 4 rS.
41 0414414w 4
0 4 0 04.5
0 4 -M5U
000
.0 A
4 8-o-s
4.4
r41w S0
V40 r lo
0000'aM C.a>
000464
4
046 01 > 0
446M% U
rH
1
.I:
0 9-4. 6
6..014
0- 00 a4 0
.404444.CA3-4440 gos~ aI- 0C640 46000 1
Q
v0- C u14 0C
44 0- V 4.) a 10 W-4
4 W 0 40-.- C
60. Q-. 44 06 -C 0S-1 0 14 06 0;00
C r4-. ; CC-4
0w 04 0w 6 0 41M 66X 44006 3 C
I6U4
U)00
13
5
4 0KC
0 C
>-14 $-.41.HOA 10
60kv0
, 40I0-01CM
C 0r 40-
>
004
004w4c4W014)
406.0.00
z 041 h r 0.4 -4 a$ 0 C ' -H 14c10
M04 4 00 > 44 44a
06 064 6 00 4 6- 6 0-444Co 04-
06 34 1100 a
600.60006 04 ,4 0. w6a6v 6 4 000 to 0-.A
0.0. 14C 14> -S-4 -- 0.C0 0, 6 1
44C - U44 u C 14
0 . %A 0; 6 0 00 0 maiCs
44- 0 N V41 >.6 0 14,- 0 4404
xw
140 N
444 MDAI
I0 10 -4 -Q -. %
6--It 004 0 006014
>- w ->0434 V-4 0a-564 00 10 440 1- C V 04J C C C04
000446 0 m54 r- 6-'~0 J0C4 oo 41 C vC00144M
-46.
m 00 -4:P
0 0. -M 4 a 40>6
U404 41V-- %w zI 00a14 6.4
6
o040C1 r ~ 1450
0>0
616
6
r4060'
Ca
C 0>.
0 64K0
'~I!
0-
r644
S 10 q00
.4004
-4
0 o
66 .0 011 8 1 4 44 0.m4 10 41C6. 03 &44
0 040 ,4
410 06 1* 604
0 v 0 0- 0 4l4-6 v040 H. 00010- 0>
6-4 0 k010 WA0- M4 - ) 014-1 .1
aa 4 6 U4444
0100 a k to 1444
C 1 0a>0 014) 1 sO4 4 .4)i 0.
c' 04
6 05- N q40a00a3 0 U C 5 N 0 1m
M' S w 1'- C 10a44 8 o,0 .4 r4 N 4 116U C 0,4 0
o600 0 a) ~-0 6 tL40, -W 0 04
14 0-H.1'--
4 0 r 0- a44 14 04 44 I2- k ) - 4
a rC - bl -P O I C 4 6
u0 041, 0.4410
41O~01 04 -- sM00-0
rem in o tH0-C 0 0 94 . >.. .4>
.wQa .C U
c'46-4.' we6 14 041
w Acm
MCIOO 1 H L 0.M 0 G4 UCE 4 4
B-4
>.4 4O 0 400 P
3000 0>14 'A
1 >0 > rI0 014J.1
w* 0 a. - I -0I 0
-A 5v
0. M
41 0 1,. -4045-
0 ow0.00. ND >
$41N r II 40 4
- I-C 0 J0Ga T %- O1 C
B-5
0 4 -$&4t
N W '"40 .$4wMN to0 %4 t
0 NO 0 0 -.4. 0 c
$
-0 400 0. 04
0 1 .44 .1 $40 0y
m0 I .. 0S$.
OO -.4 4
.- -44)0 c 4)
14 %4 A 0-A 4J A
go. 4.454 M V- V
M0 -4.0
c 80-1
1.
r -a
.. 0
0 -0.. )co
0, -c
~ 0NOO
-10-40
CO 04) 0 Ix 8t
44 4)0 .0 C$ CL
04
'4.'q0 $.F03 -4 -4 'L0 C -4
041.4 0
0.0~~
00
.00
.1a4
04
0 00 ~ 0 0' 0
0P cc40k 0
$4 0. 6-0
00 $4 10.441 .4 4rn C a4 ,
C.0 44
$4C 0 4 )00 * .. 04).00 .0 40V 04w 4J.1-0
04 w OC'
W4 a -. 40 -4 00 0 $4 .0 0 .I0
c-
1- 0 -4 010 w C -4 0O $4 >.4
V .0 0 OC0 w A a 00' %
I 0-t0o
C-'4Or..r
41 ., .4 :1 4 $4 0 $4 0143a
-1ZO
A00 0C 0-4-- 44) 0 .("4C IX 0 -'4-0
44,
U 0 .1H 0 OA 0 m 014 V
0~4
N ' H0-4 A X ) 0 0 to )
$40 0>.C-4C-C-D
4.1IA 0 4H
a.
0.
0
~M0004)1H
-Cl
02Z0 $4
00%f
CLQ 1C
0 14
000
4140
CAz 14
c
c9 y
0.0 P 4
O
k
N
40
'4
00r .0)4 to) 0.4V0
W 4)14.000
t
E0 > V( Vq 0.)
-
4 V ~ -. m. W 4
V1 U 4 000 1 04 0 0 0
H r
w14 '436
r 8 U cr0
0 0
~ C0 $
.0C -0
0 a
0 to
* v" v0 > A)1u 0
-ii
-4 0 1 040 40 V40-4 A0$
244 r~ IN) ~~4
', n -0 41
0o 4 w0
t 4-
--1034 3$
'4 0.>A
tr>1 -. u 0 $4 0-- C
c CL.- 14 0 a OU . . r
B-6
00 I I I
0, 01 10 50 0~
0
N 0O2
ow@
0.201C
4544 ~ 00. 0
'0 .. 4
-4 0a 4 5,40 '0
A ~ u 6 4
4. 00 640.
m, 54 110 Cie6..
641 04 44 0,444J4. 'A
f -.4 4j9 1.i, a.4 0 A
.. ~4 H
04
0
w
64.
co5
0 R. -a =1-p
A
U
-456. 64
0
10'
6 6N
0
0l
4
0.
6MM"
a
f
00
040m
A
> 64 06
&44 56 ~ '
0 0 00- W4-
.A a4 -0
-444~4...4 4 ~~~
-I
. 60r ~ ~ 14,1 4J'..4.
>.
~ '
A
46
a
V0.4
6 - yx a164
H4-..--
0N 0.-fm k0 '
2M 4-41J440 C'i55 a t 0 >It SV 066
I6 4-150 04..0' 0. 06 A >-I a s
CL 6 6050W
6 ~ ~ ~
006
00
~64 ~1 .4 0
0 C
W4 > 0
5 0 6-0
2054)6-0 00 4)00 63
009.a 56(. C 540 M '-40. 4'*
at
6 0 5 -40
00- -0, -t 0.4i40.4 0 00f k'- 06 04
-'6W604 .- 444 %4
40 C 00.
0l
0-100k6606 VJ r 2 1 0440a041
6 6 44 0 '0 6.
04-0 6.Oo
("40 0 Q 5k - . - k .-.. 4 4 0 0.0 4J 4
04'54*-*.o 5 14 ('4a
eq40 6
r, 00 0~'4~ 00- 66.45 00Si 0og4 1 14546
0 014 aS 4542 0 00 44 64 0'C
M0 Q00 N- 14 M 5064446
0 4 ark60.1
".4 04'5 ~ ~ ~ 045 0 ~ '
WH
0 6 0.
t-540 40-1.2
0.0
a%200;- a 0~4ev . .' 1~ 4)0#-M
*'1'4-'-.4 0 $.54 6,4 4180 -5 0~ q4 X-
60'- 04 .04 $4
CC. 54040.C c4 $4 :3
544 00 1.400 .00. 10~41 V a,0 -'w'M'c4
6c ..4 64
.c.-O ) - 5--4 -6.0 'A,0
-4250'.C. k", goA 3,2 .4 H> r. CQ $4 . 0
,a
04
0 6 e
C
0 U H4).--4 Och04 g1g 4 41
cm
404.
11.4 01 4.v04
2646260
-f =3
5406 4 6
0Q
2 Ic 0 0
4
1-I
60
C3l-4'0
00 6-5
~
64J44400 0$4604'6 6 '
*0 U0 H 4
'- : .0 1 V'5
4 4446 a. 02 0 54 4.4 -400
=1
06 -01 co to 4k -'
U-4-6 560 4 . 04 -46 0)
a -"
-- 4) ,0 v H a60 0 --0
-0440201004M442 644 -*2000!481 41 2 0x 0.>0o-4 0
UC4 4--4---
u' r 40 44*0o w -4a'
0144 1040 060 o"'05404v C104
C C4O
0 00 0
M
N0 0 C 34'A
4- CM0-.0..6.
' 54a0 ClN %A
9m-,0( V'
600 0kCa.4.-4 C 4)$ 0 N W-. - 0 )3
11.4)4
0 2 4) 0 " 4 . 0 46660r
>- : 4 .4, 0 0t
0M E4050
0 . C~ C
.. 0146
0400. > 0 0 W
6C0 00GA
E0 w4-('40 0 0 M 0C 10 05H
M L Zv 0-0. 60
Z 0 OC a6 6 046 54 0
655w 1. 66 *6 0 400- 635014'-0
'4 6 6 CO 0
4- V 4W -4 0 ' 0 A '-404 50 - 0 - -4 0 054P4.
4 14 C 10 0~~~~64
V4 C
.004 4 6 4 M. a 5 a46 4
Co%
54C V N6.4A3.44$ -4C0- p6 6 a 0
03 60 +). 664
44. 6M 0a E40 6--0P442 1. r
9. 0 0- , _
069 54 0
A 6., 446 +)5496M'.
'A'I440i'40 09
0.6 t645 "Is a
6~
'4F00is 0M - '01- 4. 0 6W.4
S1" 0545 60J aC 3I -
44.0 '6 0. 4 5 6~ -0 CA
146 0
-1 No a 5ISO
a~ jiP-n- 66
C CM CV1.0 1,-
-r Or.re 0 106 0 0--0 o40.N
0 *-AC00.Ca-11
aC a .... 6 14 - . 0 wo C0-4 4
>k4a0 -0 0 Cr
04 0o w-. ~6414644.40 k 0 > 0
C-0 0 4C 'a 0 44
_C2 Q .0 4J a0W
0 06a V W 044 6.- 4 wkC
4 4 1 ca4s
a4 i0
4&4. $-4 -. 0 =1 6 z2- .4
t46 k 3
0~
4-.
00 6z0.0444 a~~
640k*44kCO)-
0. ~
10OCOMOOO ~ ~ .0 ~
4
.~ 0 4- AO -4-
>"O 0000
t .000044.k0
0-10
-
0
44 c
!'-44 4 W0- 01 44 06.10014 k C-4 >1V~ 4)
4 M4U4VW40,r044
0. 0 XE W to.oA434 0 0 S. %60 MOr-14 C6.
4 ,0)-
6006C
V'k614w
~ 40 0 010 $40 tGo
A433k04. ,
0J
HC
CH'.4C0 -4444J
to044 0 Cf*4*4(
Id04.
0.~0.4 ma0Vto
4~Ok-4
~
4 0
14 0 040a a
06.-k.40 ZO4jI.
.43414 U0. .40 60.4 4 4U) 4 4ONO
V.4-- 6 - 0 6- 4j $4 04X~
qu k4 .4$1
0.4-
A3V~ al .owea
r W-4 0 4 1 C 6 6 1
w-44. a1030
C,44.-6- *0
li t.o r. 4o .. o-i- V V00 '-I44-0- 0~0 I 0, Cz 00- 0 1 0a40.4 4
00
so 44
a 4 61"o 0 A) c.-. 140 4Q 4 -Cn0 -4 C 00 0 6a
6~ k00 0 0 V14 a40W,4 w041 0 .0 c a .0 0 0a V C00 0
NW
k 1
00V*
C
60
63-4X.4Is6--4
-
0-00rfi
0 4J$
644.644-04-u
~U 0
-4.o0
-. 0' M
uCVON
60.C
. 0A
43 w -0k -P
11" w
00
0 6
0V
C0
C 06 0 6 M 0 C 8 0)-44044 0k
2
0.30 0044.CC 6
40 V 004-C
6C0 0-4 .
1" 00.
044to 00CO6 40 40M0A04 0-
k4
0tw 0
0 00 00 A6 00 ,2 .4
-4 r,4 0-40-
44141 ' $4C 4 V--40 0 60U .- 0§4)C403
4 -Ck 0 0 4C
14 Id.-- W, 4C V W
14 c 0 44~0) 4. 004-M
04 04 0 0I44 0
0 4 3 0 041 00h11 D,00 - >10 4 " 4C0k 0-40 0.M
to 20 W 0x 4m -4 kN-4v A 4~ W a03 VV U'.400. 0
w 6 >* X434J M-.A
do4.0
t0
1)0 ' - 4 i 0 4W4 60 41 0k 0 0 k (A ONV)0 44i) 43 4 00O0-k4 4-4 4 4- 4 O4 0 -. CO 40
C 6 60
) ak.44 0 Ch '4 0'K 4 C"0 O 0 A c
10.-(a 0W -6V2 3 0 WJO00a 4-.k0 0w 4-00 0 004 -4 44C r. -4 V0k. WO I0
, 00 3 14r. 0 3 .1 C W 0 0.40C 0t 4. 0
V.. k64..I0 0 >0 U V44 334-4 00 W'K4 0.4 6 043 w
4j 4' N .-44C1 4U)WC4 44 04-4 44 0 0 -. 4kA4d 40r 6 (n0 00.-m W
-40C ON 00 W-43.V00-0 >-~' V- 010'
kOV 0 3 400000 . .
4j% 4 , . 2-4a0r0 H V401 0 Ok'.0. 0 a 4 ' W44. 0 0.00.0
W
0 0 404ul 4 No 0 > 0)a14 0 w4-404 0
V).0 61 0r'tV4 VM 1.a. 0 0 ) R-4 0k00C4.4 V C 0: 404 *-46 ,
F3 1Ca4'-1
C40 kOC 6 0 C404.C46,00 *.04 IA4 6 H0V444 44.43 0 04 044.8 * 0o0iA
('0444 r-.-4 C -40"4 - a. N 04 0 104n a-$440IO C
W.E4 00 4 r0 00o 0 14 >4- 04 14 V 0 0 .10 4V 60 V 0.a.
1%r ) $0U941 0..-4~~ C-.4 0VC43H.4, 0.0. 4)O"
44 $~4 00 0 V4k r. 0
Ln142
w4 ~0 V
-. *0~
0
40l3
- 000 *4.4
0
z4400.
04.
to
0,
toO
-W0a
WVS
.-
k O
14
N 0034kO
4).-44 0 k 9C
4Oo W-11
0--
4
CA
X
0 44 a
0 C-4 3 O 060-44J 0
43. 0 k"4j
-144 00
44430C4 .0-
6 04 -.
0) 4444 a. c 0- 0 *6
w4 -4-0 Ck)0 0 4 C0 .
-' 014 .4 0U W w 6
021 0 0-404' -4a C 0I4-4 *.42m
00 614 10
n04 .O 0 14C - 00j40 CL- a- 0 Q 0 %4
.00
N10*- 10-l D .4 0
k LO. to 0aE-.. 4 o"-. 0 o 0 -6 30
gor w44k~ 0 V 140 W'4. V 41 .00 V W
4Jfl~0% 40 IOU we 4043 M WA4 . NO 4 0 ,
0 C4'-Q' 4 04"a a VJV
z014430 Qa;6)6r4 100' . W' 0 s-I 004 *0rCA U00U)4-4.H0
.4
C3?.0 4
~~ ~ ~
00
W-.- 6
% ..
10$~k..k.4.O0
>IN--40-4 0 00-0
rk
l 4j c~ k-401-H40-4
4 (3~-
440
fk4) 04 -401
0 01444 3-
t)oO a.C - CV.
0-.0k4
00
4-
-- 4-4 0 64
00-.44AV44I-.VA3000
04 100Z a04 00
M-4 0 0 1
-00
E. M (D c0 04-.
U)W.-414~~~~~
4 644
4 r J 444.
0~4.
V
0 tjN0>
400 ~ )-.C
0 44
4 6)
41 kOHC6444.~~ ~ ~
40400;-
*4
0k 00400-0104V
0-'-.4 0 4C 00.4C0
0.040OOCI 14 440.44 006-4 w10- 6Oq > -M-440Q6 Vk14 C C 00043 463
w - V U 43 .U0k0 4j0 0 0 k-0 4.- 10)-10004 - 04C4 00 C $4 VC
*- $ 04 10O~C
CO0 - 4 -A , r60
-14 .C60 0 -106 .4 M 0 i
sm .C O0 V OR Nr-- 40C0$44-a) 910 Z-0 -40 014-.to -460c 0 W V- 4VaV
43 04. 10CQ 4 r A 6 MW &14 H rA C-414 CD414X '.4 a0 C:Vk4 C
14 A '.H -4 0 U)-4 . A U- 06 4 aV C D U04) f 0JX$
Z 6C9 NO k-4 o A06k 430 V4z9 i >4 0U0 A 4 00.
4
r. C 10 CL 04 - 40-.' .4 60W41 3 a- 44-3q-'V. 4 4-.. 4VCK06 M
z<0 00 140I 000 a
04330- 00 CV14~ 4 4 A 0-I 4 C
v4' *4 40r.-i 43 V A404'. aC'.-. 44 !V- q a &OVr
04*- -M
=33o~'-i~ 44
W C 0MCCto 0
w A 01 4jM 0 -M 0 'A 0'A a 2; 0 n 4 30 0 4 aC'4) 44
44 .r: l0 C4JV) 40 0U 4.00-w400 60 -9. 04 4 > 0 60 M-
0 'n
6) 0 CDVCN 0-. 1 , 4)0 , a.44 c v-4 4 .00440k 4 E
C,'o > 00&U- U4- 04j Z)W0V0.CV4>30.'4.C1S04M0"3
C
43 14U)C- , 0 3 0 C V06
0 0 00 000 0cc4 C >
C H U -4.M4r COW 0.00. C16-.-4C . d an H 4H CO -4 -01
0
144 $ H a- 11'G
C,'aN . toW 0(4 3 o 0 0 r.o&104<44 WY - 04J 0 toM..04 M0046 -4C ) %
M10 j 1 4 1 0. toCA- 11 014 .0I 00a q cQ 6 -a 0 > 0300-.V -1 0W 0
40-1 4000 0 04 1w66)w H4 40 m2 04 as 4 4W>O 0
140 -0 0 . 6K0'000-604)-.-E0
,0860-AC a$ 0 -44x0 3-4 CD Z 0A 3u = c a 40 >r4
: C- 0 0 44~C 0 C41X >'-.0 0N 0p 0N00 -40
E.0 a01
6-4O-4 VWL 4o Aa J E t 4 a 4 040 ""C x Z 0A
04.4
.4C4 0 44 40410 90 04 0V "A 0
Cc a4 4
41
c "40~~040Mo0 -4 At
04 .4
0. S. 0 144
.r 0 0.444
'At - c.- .. 00
44~~~000. g444 4 *
g
5 0 -.
44
O.
C-SOC
IV"
m
90
.0
0-45Z01
00.V
11.0i
0
O0
I4Sf
0~
002,0
o 410
''UI
0.40
0
.0
140 WS
c
0 a'44 0 . U
C-.
C5 k4. 0
0
m
'.0
%0 I
:zN
k;j; IC.0r.4
V>.
C1
01
00
0 14
44 1,
4400
: 4m- tpfl A 80
4j049
0 r0.4.4.
1
0u~
~0
0444.g44
-.>00"40
0A
4.CW
k4-4A00-
.14
0 %4 '0i
x 101k
410J4C
00
r4 8930
040I
4404.
0k %4 1
-A0 a.044a.
00000004 -A
c 0 00 -4011 14 -41 C
0C: 0 4041 0 144~00
04
%44Cm,
04)4.O ~ 0- 11;
4>4
4 *U 0 o0 k 04J rS4C -
10 N O- N0 4 _ CI" 44
0
00 ; 002
4 al~
.4J 5>
N
84
HO 0
.04J
0.
V1
mW
-A0 0 14
-
0 -M>0
0
0
V.04.1
4 4a
01
-41 0 C
C'.
0C
a
44 0 C4400,400
.. 4 0 3
1"4 aa.qk
.4 0 0a 4 goo
4 >M0 10 0 0. 4. A t 40 .0C 0 g4C -00c
000~~04 0.'4
084 - 0w0kQ
444
m>1m 0 vtoV
4
I = H0-
to 0 44)& CO 0.00 ,00Ho 0 %440. - 2- 04 4) a
Co 00 4
40
V-4-
H 1-40. V0C004444V),- 0.
-A C 6 0.04
cc
-4.4
4,0 k m 0
-. 0 C6 1 VCA,0
f -.4
*4*004
0 a.,
Q1f4-.q-o
CC040
40
004so
4 0C
440
P
M40 .C 0 01- 1% -M"a P 1
>
4)0 404
040 40
0 0
~
Co -.4HkO-
4 NO-.-0 -M0OC W .14 4V43t0-i444 000.O01 04444 0. 4 *-'0>
a.4f0:-- 100 0 $ 4 0444-4N440
~4) 14 4 r4 $ C 4a 0A 4 , '- 4 41~ 4 jq
r.0 % 004 oo 1 04
V V 0, - 4 C.
at CV4 44-40 04.
m a0 0 'N - 3,toM 4 r 0 0 V440 4020 =A 0
C".4 m 41 0100.0-.'
eq00P 0 4 0- 40 4 0 >.0N444 $44 = 0c r4ON 009 .o0k
4 r r 1 OC-40.0 .404444
k _ )I 0 0 -00 %4 44 C3'4m0
4
-4$ NON0r 41 410O 440A 044o 0 k4 40C E-444-0 "qq k r
C--0... 4
0--4)1NV' Aa
- 0 ONCw0
a VCO44r. .. '. V '4 -41 00. -GO' a
a0 0 0 0 M00 .1
P 44N 0" H V 044k C 4 04 No
44
)..O-0
44030
C om0 0.41 0 *. g C On
i*. V *40
v)0.4.- 1D o44000. 0 04J4 CC1a U'40 4 Ac
0 0
.G 0 10 4.4
00.000_ -.j 4J
44
4J0N-.4
04
H
C0. 0 000
)0
r-4 a 8 '>
-OCU
H 4 MO o 14 0 0$ ._
C
0~~~~ ~ ~~
CN%4
a- ocC
CI W1
rk 10001 3 4
00301 0 0 4
H~ 43440. 414404
F 4 49 c
MM
4 0 9
X40-444-.4
14 w v
Q04 40 0 ME0 - 1) C00r
0 c 2c 9O0a 4 ,a00030 04 0 C A000.4402a 044-2 -00 0
4 Q4 C-0 00 Hv0 0
..0 z I"4Z
-s r, C 0O a 44..41- 0 4 040 440044000 0
- $0) k 4UM 0 4) 8- 0 V 004 . 0 414 0 0Q.H04 k1
mt4O
"m34 M 4 0 $4 - 0 0 40k 4.3 4 .4I
4 90 C00(a >H. 0 04MA4
4) 06E- C m4 0 r 1 - 0 '
In -P41 - 0 44 0 00 -4 *to
U
a.A-H0 4 140
%0 84
Cq 0 '04-i WO4 0 -4 4 44
4 C -S 0 ri 044-s 40 0- 0 " 040k a4.00 14 4 0i r
Ic- 0. "40 4 0 so
m W~~ go444 0' z 30 r .t"C
>. 0 x~
~
0, o4) E4a, 0Ic -0 0- M v mk j0=
14i1440M
4440
44ON 0 4 tsO v 0
AWA- a0
00
V_
c q 1
E-(Y0 0 0% 4 4a 00 4 4 0 >C0.
40A440v W40 >
fC
K 04 f0 0._00'XQ3 r0 jU
44U 0 0. 00 4 0 04
0
44
-A N .44
- 00
4 00.
a
44 0
0
0 04
"4
044 C
"U 04.
S4
*
0
4.
r.4 Q
0
0
UN
04 4 C00 0>
0 , 44
004.a4 0-t 40 'aN44 0 4 44 .. 041
040
a 4-k 0 v 406 v00 0 044v 0 - 4 r
0 40 4j 0)00 00 r. 1 040O Va 4 0 01-44
H-5 'at
0. -4v0440
00
00
.-
1 ,4 0.a 'L
0 x
00
.
",a-4-
0-V4
44
0
r
11 .
Sr 0 m4 t's44 r.-$
14 ,j I p to Q 0'm.44
M0
-'4060M0
t
04-
04
V491, A444
Il N 00
5
a
4(A
'540 to
41C o O0 o0.
0,4 -.
01
1
44r4
440
>14 00
440
W44 ~
- MN
4 0
M 4' q0
30.8
a043 u a k4Cc
"40
to'.
0
00
0 44 V 04
4u 0
0 to
.>-
010a0
Z 0 L.
01
- C -4 V44'C 0 .0 ft k 00
456 ka
0 M. 0 14 0 k 4,a444V0 -4040 "4 as5'0
41.-SC .0 C 9
4C0 00 001444toQ0 C 15)4C0
%4 0 40 Mo. 4 %'0 004
NO-v
4Cq.
0-fS W
-SOON
or O
044
soa 0
C
44
4
4 N
40
44 41 J
ty0-
~ 4
04
020 0 4 9 4.
040 M 04k. 0
4r.5-
00J 0. Co-s 041 9 40a
4..4C
v-. a 0
k*M.4
00 4.4 4 4 4 43 au * C3
31041@ 0 >
I'll, I Me
I~4' 0 (A~44
!' ., m ,. . j " -0 AlO O400 C V 00N.
gn.' goo 0 0U44--
4- 040
A 9M - 44 .1 OM 0 0 1 1
0Mg0 4 00 0
0. 1. 09-' 0- ".041 C O
4 0 au 440 04 0
44 A4 000" 0 0 - j a4
810 m o A a 1.0 0 0 4 4 k C u3 , -
U .1 .)0 u aIj j &~~V4~4
0 a4 4 1
B-9
M-0 0 -I.V0'.4J
40f -Or 4X44-4-lIEU1
~~~
El-- SO
ON~
.4M-C
O
x
*.-.0
~~~
$40)I
0S
$4 -- ri
1440M
40 W
0 a.CM 0O4114 C
X0j1044.4
.4
. 0 0
$40 00I
044 j 41 a a)S.0.. aI 44 Ue 04 - N 0 41 *$4
0.40
'N a4U~ I
in03$ ouE$44)
o C,99aI- t 041 4) -0 a A uUCS04
01
0C4.-44 ' 41) 0- 0 -4 Mo 4)
0, ~ 0 C.0 4 aECa
04 14CZ
M 0C
5 0 a4) t 4 00
41%-* 5
00 r
0'4
C041
0 V -1kU
V a 0N
o
4JI54
SC to
'4j 0 L
V
gE0
04.
~0v
-A -A a
-00.45 4 4 $4 I 0 O0
00 04 0-4C4 w 4 0 650 0 .0 ' >..CC
O a0.4 a a k 0% 46.
00 0 w
4) 044 410 0W 4) 0.4x0 W
41 C 0 ' -A445
0
0-
0'- I 4
~ ~ a000 4C4$C
0 1040
do*
E 0 0-4
0
a
u
a-00-
VO0 gr- 0 40.
4
C
co'300: -I40 04> 0 04- $44V04 0$
"4$4~ 0
$4.4 .44, ) ..U -0C.00 04H 06 .0 04 2"40
40 004 U 0 -44 .4-k4041 ot
S. 44 > 0U 4) E.-I .U I IU.C.1
1it .100404.0 14
.0000 0U$400'5
r.I 1 0'- C
V.- 0'.
, . v .0C-Hog-0' $41E CO
-. 10 V0 a 0 a.1. S4a $40 0, 0.4UOO CO 0 J)4 w
O 01 4 , 0E. 4jU,EU
C''.0 rQ 00 Ar. fMo $41 a A
a
050.4.00'a o4 40 0- 0 0. W -MU4 4 )V4
00 p4. A5 0C I0 E. 0 0 4
o0 - CE1"
w' .~ 044 0 0 0 44H 44
0 -)- U0. M. 4) .0 0 0 4pEA :04 0.
X E : -4J
14M>..0
V US00N$
4 0 u &0 $4,01 5.
.0 0 (. 04C4 040CS 0 % -4 v
.4 04,4 54
*4.10044k n. a$ 4 4 0>$4 ( $
$4MO VOUq~~> - 0C
4A0 -'D5 $44 a .4 $~-00I. a00M40 W. $4
$NA *-c44 -H $4. 0) go0 . C 004 01 0$V .
44.-I *0 rE $4 4 0 A-4 0 10 0 0. moCS C '0o; 4 Ac 0
CM O1 .0, -4 M 40 4010 400 Cr$. $40 1 - [ 0w Go CU8
,nC4 a j- 0 $004~ 0 G.400 0 0 C1CS
0 to0 3C . ic C' aI~4400O to00 4 h41 44 0.0 , 0, 00
U C
r w
NCrI$4 o1104, 4.1)4) r44 0 0 rI 0.4 $4 J
I 4% N 0-.4.0 0 000 0 0 4 40 W 44J0 40 *. 4 1 505
40$4fl-4
als-. oW0 -
004)CI MO4- VC V4%40-d
UV 0 .4) CS
L
-A)0 U24 W44 1 5.0.0
00a..
4
(A0 VQ 0 H ON0k444 0 f 0 -- - 00C $04o.1 W 0
t -4 V
.AM 0 A 401W $0U 0. 0U0 C CE r
00 50 C.
]
04 a- 0am 01 0 .04tE a :0 00k
Un >-w.- CC.4C.-IC440It0EC C 44V 4 .4Hw.-S CC 40 -
1445 -0 DN w to.04 0EA0.44 0C u240O 301 U04.1V000
C40'b . 0 to1 0 A0 C. CA'a0
0 00 0 $404 00 a
0 V as . 0 5>. 40. 0 uy04 SE o $4 02 0.4 H0 9C
J 4.)0I
z k-403-.4 4 $0 . CC1
E$ 44 $4W0. C4 44 E44 A 3. 40
NUN'.400 N$ 0C04 0V. Vp
Pk00 0C4 %U%0 00.0 V 50.0
0 zN)
ty.4-I--0
.
N
4U00.2~.-I0O'P)
4c
4 a
04
E
00 ' Mo.44-
4. 1 - 4).C
; 14$40 -
r a
4$'a 44
4 $-4 104
4 $4$4
>.a4 0
N4Il. 4G-0
0-z P a
$O40)
041-4
'00
00.0or
4OZUUV)0
.0.4O
C-4.
4
aE .
4)
.C)00
0004-4 0
40 .0 V 0.
a4-4 a$...IZ 9 .0 0.01 to 0 .4
c0 4.4$000_J4'.
01 0 w 111 1 . 4E)04) 0.o0--MC0 0
0,
I.4MC r4 0EA0 MOEES4EU4
04.>~rf- -10
0.0C040 000.M0a) $4C4
E-.0 '-4 ~~~
$4$001444.444 04-0 04000
$44t
a' 404
U.4-4 01 0'w0
.I'If4.
4.p4
4004)r.W
U0
0
a'
Q4 r00M 404JA4
4) 40
0
5405N 0V0.I4
M4
4) $
4_0
4
A
4 . 0 04400404
A* 00 C 00C .00-I> n q00400C 0 4 0
0 ) 0 400I C-)
4) ;:.0 9-4))0 U0414 0%4Z S.0
04 0 M 4A '0 UO.I0)-I0
0 00 -)4 *A
.44J.0f4OU 0 0444)0
$4 $4 0E %.E 0 0 -4 .-40 4 0
00 M .00?SO
-
S-IMCS- M'0 IQ a SP0O44h 4 0 v0 .4.a
04.
U .40 4 O4 4 ) 0 . 1 2 444 w 44C.W4 0 00CO.
.4 *~4 .4-C $.U 4 -4 4)0 104V0 v UO0
V) .0 0 Q.* 4) %04.1C40r 0 00 U001 4J. 0 0000 .4t
0 *,4
en0-$4 0).--A ) 0 - 4 )0 4441 00. .0 C44I
V.4aso~ ~0
a00v .M~ 0.4 4J 4 * 41-440
0 J.C
4I '. 413.8 0. 001 ) r In t 0 4$4 4) IV S4
3 4 0 -. 40 4).0 A W $4 40 000 4C . $
00 00 4 r E v 0 $4.a.Cw -0 .4 04 11
Ok$4
& M
0
~0"E
.
0...
04
44)0
0-
s x 0
0$4*. &
00 4
)0 5
14 0 .
40
'00
00444
0 'j
00- V 004 -. 00 %444 500 1 0.4 .0 0
0 00 U V0IV O
04 0'0-I
m a C 09 0 P3UV5 401 V JV4$ A 444)
40 k.C.f 0.0 14 4-100 k 04 vow Co 00 0 4) 04)UI
4001 401At4 a a V0 4 4454 C. w4
-.- 0 a 0 q 4
N$00 W0V0 A040 In 10 0 1 0A %
400 000"4 %4.1; a4 -S .
00 0 4).04.1 0 0 o0 -40 c. ou'
M, - 0 . 0%q C)4 RAN.4 C .
$419
",0
aori 31In$ A r 4 4
--. -1 ; 'p. .i
-: ~~ ~
4M ~ ~ 0r~. 4(
.-
1 a U 1 10~ '. 1 :1 "1~.
1)30 40 4 0 ov.
44'l4 1 q 14) IS44a4 14 04 W 0
2.- 0
44 N
01 .- 0
.. 1~~~
Jj 3C0 .
6 41
14 N
O 6.-44 04.e-
V USC
t3 10
0 C O
4
0r410. n oo m 6 06 r0
za -o ,$
6*
1 8 t4'e 606V Oa o1 -a0a.".
-
4 C 6% % .44- 1* 4 04 0 V -4 0 4'4
V O01 " 40 I C
el 01: .a:. 611 .1 4' J ' C
I-1 r. 0 S O 4' 16 Q
1-.0 . 444o-'v0C
0,00, 0 4'. 4' k a ,- -01VC
at ;C66IS 0 -4 .-I
8A.
0 ol' IF'0 4t4 010a$ 4 2r C 0 p4' 0 r
C~~~~1 -00 OR'1 kn44 - .. V
00 2 OO
aC 0x . 016> mC 0 04 O $4 0 C AO'1 '4 U %
414 W
6A6A -4 6-4 C 0 114 0 - -OC.00
14
O' CO 0C -4.0 ~ 4V 1
31 0 W A~
0.v - MA
o10tI
V-4
C
-0 -, N2 C
~
0, 0 .IN -14CC 'C . C0
OMCH 14 .4 0. I0 Ad 4,46 .01j
0
~00 81CWC-I0
0OO 4 C 4C6 0 64 0414
$,C 16-.4 00m COo-
0 k 140 m 6 0 .0 k
mM CC >OC
40 1
Cr.U'W 0 -4
' 0,6 00 01 WV V U>. -, 6-
-4r4 O ' . I- - 0 4 4 4 C g o;V-
O C -I O4 140 I
~M% - 6 6 0 0 .0 45 M l V . . C- If6
400~ CWco4' C N.0
Lo,1 624
(f.*.1(-at -4. 0A >OQ-- COO
i10 v 14- 10 r. 1 11 .00-g--A' 1-4 Is' 0
9. 9L 4 . ->0U4 0 14 4 C4 'U CIO r 0 Co. W.6 C
0 011
>
m
1- 0.0Q61
CCbl 4o04 0
g 010.0.
0 .0
~'44~4
00
-t
t 'W"
C0 cO P 0
-400
&-'
4.
01414-S4a0' 14V
wo-.46. 4
6
6.0r
414
A - 04'
040 - 4'6 .04 0 A 0 %4 ' VMC 0.b > 64 6
C 4Os-06446 4 a.r 0 >,c 0 0a1 0 C
toC
$-414 C4 0 C 040 2 6~ -04 0-4 so.06
C
.420 0 S4 t.I 0- W4'1 vW04 C6 C -
'40~. 14000 4 VV A646 v4.I 1 24
:1 0 C140
>0"0 6O1.) 4 COVO 00.. 0.-C v%0 '' so40H
V 14 slo60
4 - 6 ,O% ,- ag
000 . 0% ' 0 N f 1 Z0 60 . 9 1 1-40 MOC 0a4Z60
'00 O
0 4UI. W I4' 0 -4 60 0 0 0 . 0 - 0 O - V0C
' -A--- 0 40 .4O0 O 4' N
0V H 0V 000CC 0 4 V 4 CVC
4.'6 0
u v 0% >.-
oW1 A a MW .4I. 0 -4 04", - 44 0 .C 04
01 1oC 01W
. _841 4. 4 . 0 C40C444V 0 64
U-
1 a C a 0. 0 40 .. 0.C0 o 6- T'.
u o S>V06 00 ..4 0 a 8 --
AVOO 1- 0 0 S
-4N00 a04401 '~ 44I
*M 6OW
co.- 00,40.0 4 9.0 0 a a W4.O V26
1%c INO1V 0 0 6M 0 41 4' 1,1 1
14 0 IA4 0m0 . 4 V-4;41V14 0 -4 4 9n 4O tP .
0~~~~~~0
-40C.1600 C OCM *S00' 14 . A V~f0
4 40400- 40 -4
R4C 0-W u m 06HUM01 A 00V .4 0 $4 $64 0
14gEC
06.44
-4 60. OO 0O061 C 44
01O~ all 00 C 1
0
.
~N
.06 4
0 .
0% O 01 0 61.640 k4
E CYOa.
' H 0 v >O IV10
me-
0~~~~~~~~~~
0M1 O -
10001*
4 0 P.I 0V. A
6 0"%W.-4
iC 0 A ..4
w8W.W.41.-
.C. 4 v4 0
1.4464 1 -4010V
,
.0.U4 .1 0
mw 0 6 6 0 6 044 V01 14. -0 - 0>0 1 014 4 0
46344 0MO 00.4'I 0-M. 440 4 64 M-4 6C60%
4M %
CC 44
r-4 $4-M 0 a00
VOW P-004-0 MOO'14 1 t o C0 0s c " 0
6-0.-4 '-4 .4 M0
0. 4Ov D 4 0m4 0I4440
0N
C, k 1 -3* 06
0~410.0 C-OkJ 0 OC 4 0 00-4.0-
0
0.0040v 4 4 34 8 0 0 %04 C OO 8 43
0
0 04
6J 0.-OC
,0w6
a~0V.0)40
v 04 0 044.0 140
0-a- I, Q4 *~-II043. 0 00 4
k ) . 144.a
0C4 .0.C to04
0 0 000 U)JU
40. X4.0 V.04. 0 V,4
41C4. 060
03
C
00
0U C V44r-4
44 3Ca at
0
43
.04-
C 44430
CU).
04- C0044
0044 -V4. C.r-
09.M C C-0-b1>. 0o
C 443 0 014, 04 4 V40-C04.0 0 z 040, >0
0 e 0(.4 3 -4 00 04C 0 .0. 04I 0U).0C-4
N 0 00 4 IM
*4 4$4v 0 40 40 4 3 04. a C 44*
.C6
40o4.0 )4O44 vC44 40 40 443D -.. 0. kW . .0 a'4
-4-1w C x0004.0
0
C6
000 3( t0
40
cc 0C 4.
5. . 0
4so4
40
4 . 0 0 C>0 4. P,33
0. 4-j.13N4
0.4 Ja 00
C. * a 044
g 0
- (00
34.- a 0
0 1 .
c c.
*4 C4
a
.4 0 4
z .c >.4 .0 0 0 44 4 0
9 06Io000 4 04 C04.0 60 - 0e014 V04 k 0u v C 0. 044
-~~404( 44.) CO U) >. C..0
0 0 C. 0
V.00- 0
3CZO 0 00z 4 4 4 00C 304 0M% 0. C 4 . . 0 . 4 4.,.
M4
.4
0D0 ,
v0
.4- 4V-
0
~00.4
-4 . 0>.a.44440 0 * to
-
-00
k
0-
0C-- 1
0.M
>0
*.40ACm6 0
.
.40- 0
40
44
6j 0- 0a0. 4 04 C .40
01CCV 0-4C4 C U040 C
w4 - 04 0 4 4>U -4400 4 0 4j g 34.44
00 04 644 0 C4I
A. 4 4 0- Ci 0' 0 0 k I a 41 0 w 00 00
0 04V 1 4 U09%4 09 00 -1 . 0 014S4 g5) 4 0 43 C .4 0
"J4 4 -4 004 0-0-.04 4.,0 04 504 4 1344
4I4450.00
C- -44
.- fl0.4.
a. M SC
-a,-440
4 00044
43 0
4 0.-44
0F 4 z0 00 .Os )- O.004 00
4 M
044C 00.00000k0
>0 ON
0 411-400
0 434443C000
0-0- .043
9 2 440
004-
-460i0 C VC - r3130414.40.>4
C -k 040 C coot C -4-4g I C
-4
CO 0 00 .-4 00-1O 4w40 V0 a0
. 40 4 0 4.0 b>, 434
0 0 0x 0U b404
OC
4 >-46 0 4 1 0 F w C aC0 - .4 -4a
0 4 .0 IC-b 44 0 4.00 4
V0044, .0 04 440
001C. ; '0 4 k u4 AS.
. 0 0 .0> 0 . 0.4440
Or--4 M00 0 0 1
00 0.CDQ
.40-.4004. .401o 44 0o 01400 041 0 0 0 a 0
.6004 a
.l C000 C6 ., 4 t)
4044 :0I * ,,!---4. 0 C0043 43 40 0 0 .0
0C 3 0to 430 0 4 0 6M0 4 04 .40 40 W.0C-
644-4 C 0C 0 a 0 C-0 W4 04 00k 0> . p. >. 0.0 a asol
. 0 4 %.0 OC.4 04 . a A0.r- - o .080.
414 I00 .
"43 g o -u 40
-. e . 4 4 0.6 4 00 Q-A 44 U.000 04 -
444J4 CU4I 444-4 04 6. 0U)4D. 0 000 - 4 0 0
4 0 4 0 6 14-4 0 ..
10
U C 404
4 -k -A 4A~ V0 Wg0 k .40
0 a-4 a w34 %~04- 40M4343044
140 V0.0 0 W.4 0. 04 4141E 0o 0 044 444 . 4 . g 0.40 0
- >4 0 00 m
40-. 0 41 a C
~kq
00 400in
a4. 041 -A 19 4a C- %>0 c.0 43440
, *o
e 36"4--
4 1
. 0 a~ 4.000~40
k
43 -443
000
r02
9
lax4
1,0 01 0 0. a840 -44 - . 44
144440
0.4
0 44
4 41 0
004
44A 04 . I v444
Z0 ; ~mC .!1 0A0
40
A~)
r 1"~. C>, 440
0~ -0000 0 4 4
I.4 -II4 0 . .4 0q 1 V) so-4a0
.40 440
0 1'm R C 0d Z-0. 4 0 N 4 0>I .0.0zw00 4 1
aUA 0 > 0 0 0> 0J-
4
ko ,0 0 .3 C 04 -U 0.4 I 444 0 I444&I>goI 0
044 0 0044 0-1 44 In- 60w 6 .'a0
1 r
4.43
3 0 0 0 1 %.I.444to C *.
a . A,0 0$ 0 $4 w. j 4 4 44- 0 O .
B- 12
.w
11
e $ In ~ 1 4 1 4C)
r4~~ 4 O.14130
1400 m 0m0
.4. ~rC48 W.d1d
4v 0
4041.
048-in.o F8 444)0 84. rv 1 0 v6
iv- 6
.4 4
e..404~r-a G-.0 *~~~0 6 6 0 . O N..~~
.~
-a 14 a46O 04~
604' Sal
*6 *0 00 0 14 .46
wop k IeU C) 0 1 600 4
w a. .A.60 t" 6 k~6 14
0 oe A*,oeva. -4~ ~ 1 'D $
348es
0.403.4
0C)
%0o
4 :4 4 14-414u fm 06.It
6440.
r- -. 44 00.
0 6'C.. 14
0,.4!g
v 0.4 1. 400ko 0 .4MA 11
'08 4 64$41
,8 A1 r.
I s -4 % Is-z4;6
% z.04 f
6. I: -X.00 .4 6. 0.01 n6*4
4 4 0 4 01 4414 Ua4844
140. .. 0-48 .. W
o~Z.. 0 414 0 0. 0 . 6--4 60
.4 4 r4 04 In '4 4
c' 0 4 0a '
- 0 .0 11 4o$0Q 4
00 0 0..4 .40 061.4V .004
$0,80 >,
:.41 0.-0 006Q4.00OC) - 14
ag 9 0ac a
a4.
H 8 c~4
4 0M <
0 rA Mo
.GC 48
00 ao 4 9G goo a ~ c
. 1mGOC - . U N $
6 0 O t* .
*..4 r8
G400 4 0 'I6 $J6~ .4 4. 06 01404' .c1 0 0 I : 4 9"14
0 0 a-~ .684 0 0
16.
Or a**
400 1
Vt0
*046041.8-
6
0 44.e-4104 w 6146of soC)0
&V
.e 4 0 *466600c1'..
.40. 'aV0 '-4 CEO -G.
0,16 10 rO 11 10
. 4 00W me_ G6 o401r4 0 .61 .
0 4a .G. 04 ow10
1, 6.o 0 $1 00;.3 >. U. 146 - 0 - .0
1 0 040 466
-4G-48M mcoG 0 =a6. 6.M4 C840
I 4 .6 IC a. 4 c4 0. 40 *ft.0tpp4 0 6 48
400 0 C00-0 41 G 4' 6 %4C 0I0
04144 4,0 1 Il4. 0",
01 o 61 m 060 i.4
Z..o
Xi.. 1
A~.6 or- . 0' 086 l 1144, 0>7
306 N, C) -481 c.. 40 . 0 u 14
64-
.q) -4.40% > 001.4 X4 U C -. 6,0' 0 -4 0 060
O 4
6-.'~G48
RIO-
0M
.41400001
0 14-
.0404
~g 0
0-
0 W N
a c .0486)4.4
-.4
000486eG C-G4
4
:.-
- o00 x
*6
4NONO.-4.~
6 . 0 0 4W'*
0 00.$_0
.4 &1
G
-M
61 a 40
*640
000 ~0 ;~.0414G2U41"-
'I.
NON
1 40G 4
0 4 h6
(A016. 80
6
6...
g
1H
-4. 6
H ol 14 '0- ON" 0' -
08-
E -3
oom0~ls
-AP4C
- 4 0-.4104 4 00 4 6 80, I
q4 4 W~ 00.' "I40414148.0
0 - 0
40.4 "
))~
a~. 113O 0040G~ ~ 4
$
6x 0'004. o c0 -3.41 0 4 0 600 0064
%4caM. OU .G2480 c 0~ 61.444
q04 8').
0 041~--9. 0 to 060 a
04-.4 4-040 6
1.a4oo000. 0 .16W6 0 jO ... r- r.46666V0
4 4-4
4 ) 6 m s0 1404 0060041G4
U4 1 iP4 00
-- 4
430U 1,
61400
6 00.-i)4 148 6- 6u 43 4 a V Sot o..4 A068 *-A0.4. a
m48.4).8 GA&0
.46
* 4 06 0i06 U
1444'.G 46
0I.
e0 OV3N4 $4G48.44
k4 S80-06 V
iG'448 4 f m 0
('Z04 -O 0 4 01M
6 W.mG
c 0 A A0 1 014C-
0do 0 4r;G2 t -4 8
a.- .4 '60 > 0. V T C 00
-A)..3 -1 I t ' r661.
4.-AG
V4 m -14t*
8
14 0.06 -06 0.1
r. "
a;0
.00
1-46 1
0v*>,q
)3..6 M
.
I 14 40,
1
IV 0%
04- 0 46
0i 0001$C4
c%
0-40
04'480 - .% 48 040-4 mm .A.41)0
644 40 4 al r). 0 c~ go q
le64814 also~
C..4 6.60 006 V06040 C 6 64K 0614
a-% -A 4
*00
0 40
10811
00
0404 6-4m'C0) 0 ~ 0- 0 , '
$.484 M. 4k48G44 C 5io 1 ",
" _a 48 0at G0-4 0- .4 0 0
60
0 m 61"401 0 44
00 484 '4
4"M
44 ..4 6 60400 8
ow~6. 06.6'4
$4 r g0
0-a 14 0 0 4 a s
In~6.
0486.-44 %
0-6 -4 n 30 to4G40 - 6 6 .
m46 0 1-4 08 6
m141 - 140- 0 0
114 9 1 :0-4
66 0 C 4-1 ; 3 o
00- k - 40 -, 40 4 4 . 0- 04go aZ l &4
.4 V
$61 *.0 r-00.-
Fc ; 83r .04 .44 a 46Ag
.4 to
04hIj~t 00
06.-~~~e
144 - 04I.L
B- 13
-40. 30 W949 Od
2,4 000 A6
- 4 9 ,O N - 6: 0 e k ..
c-a A~ . -
*.to 0 63 N04 IOcM 04
a.00 0
'461UAtr40
4
;-s~c9 -00
$40.4 % H
a1.
U C
6 f
64
0
.4
448
c 1 0 -4 0
0 1.4 14 n 0 z Ove v a 0 w
E40.4 vea
00o .0 9440fZ4"44"J
0. tl 4 1 k 0 4J: 4
0-40' *64~ .- 096 O49...
a40
~
Ij,9 94
C 104
014 49
-46bL.4
... ,
.
*0
sum 4 -6094
LO t)4 14 ' 00,W C44 0 0w14 ~ kc.60
0 0 .06 j
-1494 -
M
6
rQ 04
'44. 0-14 6C.6.040Om
94~~~.0-4 111 0-
goof.....
40 >_
010 c 0 94- 99 4 ---
6 0. .04 4
I~
0 .. g.0
a..~' 0 .64 490,a. a4..
,4~
r 10 -41 1 O v3 a
00. 0k00CA 0
w4 006146 0 0 14000490-40w
(
'oZ-96-449 4 40. 0 0 k4 w 0 .4.-40,..406-o
4 W a'4 O0>1.00 0Z9U. j a 26
-.414 4 449 c
0~ V 0 E0.M .*44', 0
,q .0' -4 O6S6-.C -00 w a00 * .4 00
111 M0 0 N ' 4 104 0. 0..
-6 0 26
1-4 _1 0 1;A -4;.- 0 0 Ma 6. VS
C44a 0 0 -4 -40 1- V=
494 00414 49--0,40
;4 10 .0 1 496
a44- 090
r 4 - 01 0 0
0 ;q 0, W 414 a A414
M.1 066c 0 061
940 04.m>0 _ 0-1 0V6a
4 14W 0
*; 0. -
94'0606400694'
0--V0.0.00E-0C#C WI'6I940a-0 2~ - g.
'460 0 .
94 0- 00 6 %4
.0 c$:1
.NN 1i0 49.4)49~~I
049 v9 P4S CL -4.9 soV404
66. 006 0
464
to . 00 W.f-.41
6,, g40.0 00944 494. - 6
1000 a
%40461400
q.*00 0.-4-l
009
066~~~~~~
6CV00wU o4. a -. 0'460
x j-4M--~a
'4 f6 9 )O.4 at* 0-.64
0 4 0-49 0A
~
P A - C (A0 60 C a- -9
c ' 0 14
U .0.-
' 4C-
E .2 94 116 Go 9 0 V.4 0- %4
E- 01 1'0. 060-449v -4 00a4Fa
. ~ . 401 6. $, 49..l..0
'4oC1 C61 Z 90 M
'A490o
C4~ C4 014af0
& 0
C
,.
t 49
.. a.
MO Vaa
* 03 .
-
V1-
4a
4
~4C000006V
-4
.4N 449 .
m4f230
6A
-40 I
84
100 66 14N
CK)4 040-0
- k
04
A k
C o 0 90 64! 10 94
40 4l 00
4
4 .409C4 Cc A, 31m
00 0a
Me 6 00-
I .06 94 W 149
49
04 000 6 U40
0.404900 a C001 H.01
14u. ca
63-49
4 ..4 14. 0490 - 94 90 -4 .44 0'44 4r- I.kx.4
C 14 140 no 049'9 lowr
Cc0 EM - 49 c a6k.4 9 4N0.4445 14 0 Nag
004- me Z'4.a4l-c0 6 A0). v 9 U4010
cc
I0 $4 s 4
0 N 3
0 w0% 00 $ '4n C 0
-409 00 0W 09
>000 Z j
~ %. 4 C, 4 004V*~04 0 0 3-1 4-00
0 tru 0
.8~~
94 p 0
4'4
4
v8 0Ik .. 3,
K~ oq 4 o -.
4 4'00 JoA 01-O
810 16696
~~41 Il 0
r 4
I~A4
P A 41 W0.0~:''
I 4~
43 0k% %4. 0
.4 AZ A se s >4 1 1 InIC
94c rI. 0w0 .In a44 0C410 .4
ZVS40 VZ4 1
01 00 0 3,- -4. . ai4
-4
FA a~
~-144 ~A tP~-4~
oa-o
4.4w1
~~~~.JQI48A 4 00 -
v111no 0a_4
0 -. 4 . 04I
B-14
90 0 tp &C"*A 0060
0112
0 ..
101,
0al 1401
a.;;S
-4
0V
~ In
I-C
0
m.6-4
4
C"
a~~~4 -4 , I~u1, .
.~t.1 .~).c
C10 6 1a 14
61V 6 14 00
60 44
14 4-0 6C 10046 -4 5 0A411 a'
6 OC
a-oo 96* a 4J r-061
"4C~ -441 56 -.
UA '.- 6 2 a6 C~~~ 144.444 4
5$C)4
u 0.
6 no4-
00
QCU)I ~
a0.0a n
04 0 11046
4040
.1 1644a .4-.1
V6-
k 0' 0! 0 44 I 10 -11.4 a.IS%04449 car0 -4,0 ;i48
.4VC In 06 416C CS-k"4
604 In~~
0
.4 0
m N
0.~c 6
V
660-4
404Y0 0
C
-0 146.2
04604
~ In466
. 9SR401 a-A
a.6~4 0 0
q64141
04
40
0.V 06 00 -0. 06.41 0 0 4-I
0 I'l 6..6N
140s4 e
06 W.44~
.- 0 0 0.
44 )0 v 01 01& -4 c 614 06 qu441-
14 .% 4 *0 0 1
44 r > C 1 4 C 6 r006 - 014-4 E464 -,- 06 41 1f
6-.$-4 C.VV460 64 aC 0I4C I'0
6--.- :6C-'---614
rNj0 40 6C 06 . No 4 "-r CD0 >'. 4 0 >.140.6-
v g1. 4 &, ,4 *6 141,A%4m0o %C
Ma4> 0 0 400 - 60
aa 4 £6Iw 41 V inC V464 0- 106-4
61O-3 * 4 10O 441re c S 604 -00a 0.6 > &, 0 4.
k
.qU) .
0614fl'5-0
3 1h4 ii
C9t;.414
-.
-. ~.0.'a4
C
66
- 0
0142
1,' 0 L 0 0041 C4 a 4-,1;4
100.0 - IL
41 ON 0 6604 C0 OV0c600 0 0 E , O 6 6 C 60
14 6a 06
t'"S 0%41 11 I "-C
.4 I411.4
C0 a4C60 > a 14qC
-40. '04.4 6 > . 6
Q$4 0 03v-
- c CO .1.- 06
404-a a6 i c1-.
1 4 0 0
z V-4 0 4 V 1-414
4, 6 .
~~.-44~$ C
*EM-
410
6416
--
1 a - ,46
V
1-1090 111
-. 0 1 a'
66 4
14CC) 064
E-4.4 In 64 4 $.I%.
1 64.4V. A 64 6C1 P4 -4I '4 k 410 c t C 061 6
4j
X~
40
~ 64
-400
C%
kI 4 1
- 1441*01"~ ; '
6
N
01 ' 6 '4-
W )-, C-1 0a .4 - 4 c-661
1
1 . In4 4M 21 -. tI. O42-2 oz l
146)'044 k 10
CO -O 6 6 66 0NC304 0. IVC6Cv-r.4- -Cw >-j0 0-46
114 16 Vp 4.
OC'- V1 14 $4C 0V0 410.4 . .z
No 1C 4ci ) 0 1 a41
.0
1
0 . .- 14 6,
16 -. C1 140 600
a- po41 C .U vO 41w
V1 -4 en0.) -Ac
60 .. 4 000
U.~ A .0 a'-cc.,
06 -4 4 14 I4 44 N0
..1I0 4 0040W0
0. 6 .. j 6 C 44P
t0
N440 co 6 -4 -0-- X4 - 65 4 "1 2 -
41.456 a -~j 6 v~
41 41.4 S r64 .0 6
6~ -0 6041 -4 .. 41
0C *
0I
14-w4 a
-Po
a. S
.41- . 0
-
6141q0 14 Q0 1 o 0-6-4 j 6
0-0 6; In 6- "C'm~n1 aC jIso. , 4 I
r140 -. 0 96 V4rn6
V~*6
j 4:
$4 j k. 0 .0 0
t 1 410 0J 4,4 I cc 9%04 0
o q04:I!,-. 3.f I
6046 o j ,41-
ipI41r 004A.Rk'o .I I1 V, 5 - 4
0 -4
h- 4 B3 -A c S I1': 4
1 a01> '3415
" .. 0 0 a~
aj t40o0
0 1 41 0 v1 6
0a
64 1100 .N00 -.11.11Oo~-'0 v0 4 0 0 -I . 6 6
a4 0 p41
p4-0.c s o a 0.1
54 ~ 66 .14 0
0
U201j
'- .4
t38 .! 02-410
zoe 50l'
0 006lua a .I c. -m 4354-
o44s
a- .0 a21
-.44 044 1U 4 6 64 4 0t0
-. 4 6-00 041 0 0 OW
14 a~- .02 0 4 -6o 0 - ' . . . 1V.i
22
60 6410 OW
6Q4 :. 1;. 0 -16 066040
oz 0 -4
601
0a1 0.'
05 44
v40
Q
0254 -'004141V 0 0 41 00 41 005 5 2 5 0 4 1
4*0
a g. 0 0~-JA 4 x Vwai . 0w- 000 6 O
xx4 .10400040C
3 1 . toA.01.
'4 ..
a..4 44 , %c ...4 04 !
c 00 41 MRA 0-
B- 16
0
-
1 M4.
I0
54V 00 4 6 CM
0 514
ak .4
6
%4
0 A
- tp
G0'Vaa
4 )64)
0 4a46
t 44
51acc0
46*4 44 0654 a
'46 4.1ov a044
54 *1gg 04 4)0J
6-4UN 4 4j
60. 0 1
~so Vu1- 0-1A CA as. ~ iS 0'4
V4A a 0. 0
S -4 .004910)4
"A0 660 . a~5 1
is%d ? M a so .0 .,.r FN 0
4j
0' 40
0 C04W4
0.1
E
~0
4 .6
6-45 46
))
4.4 4A a~
P13 a.4~
**.. 4)6 64 0.
low0S
H>0 -4 -4 0 >,0 o.14 0 -,a
-.4 60V4 OS & 14' a 4-6. 0% ty. 0649-0466 . I f0
0 90 > 1o
I446 4 %4-6.
6 0 6 0 ~4)-
:0.20 1 low~
me
-.44) 056
OM06
4) 4-
04404
~6 6-'
0
1~
0 46
4
10
0146r
c1.0 - 645
64-j.0
-4
.4 -5.9
p 16 6 .- 0 >))1
64 P3.
0%A -.. 40CA
05146~ 1.-.4 04 6- c- - 4 2 0 104
03 6,00
4 0 4 N4 A4.4 4 V
V44.- cm x)64A.
0 0
'4
U1
I N4 00
J
' 0
0-~4
E00
-(*104)
04) ~ 4
64...S 49.
0-0 GA'- so; 4*4 C040 04
04
a3t1 406 0 04J 0 0 C54 -. 44)04'-4 0040
0>- -4 4 --- 49 )2 64 6w 0..00 4 5.n'-. 4)4)
4) 066W -4 '-. 6->' 446 -0a * 4.4.46i144
01 00 *.-40 0"Z
%-4-4 0. 0
2444. 10le6a9
So14
0
010 2 -4 3-. A4 4"
-.4 6 6A.
0 6 0 0
515404 - 6 64
6 ,)0-060 a.4
01I 360 " 01 00 01
t . 4 $ 4 0 02 so 601 66
Q. -0 C-1- 66
140.0 A*010V0' 0 M.' "I 'r63'a 06
04- 4j@-. 0 4) . 44.-l 056 04 c .04) 64j
3--60 0 00 04Jm6 .- 46-.4 266HC CL%60'64).-.
.0 4 0 6t $3- 1 In M1 06so -4 6.0 60450 4
31-44)-4 -4. Q0. 414 1 60 a. a-66 6 01t S 6000 661
0
4)0
.
535 2
01 0
41
.14
V
04-4
00
N
4). X
A
aG
0-4NO.541-45
0660496 54H0
04J4404a0NV1-
14
$4-4 040 4 OA0'
A-. 06
44-40-66 V1 64) .4-45 -x
-. 654 01 4 30 V rL. ).0 66 84~410i
V1.
F8-4
6
0' -0 V 5 0 6 03 0 0 u--5' 0 046.0 14.-4C
A VC 05= 040.5.0 5A 0 6 .- 0 R 0
ocM 0 4j4)
4 $O%4-4
6 -4M 4 r 141440AC NO 410 -4 94 6 6664 0004)
'-
OD46 ;4r)4)§40
I I0.--0% Go. 6614
140-
60 .4to
.-43- .I
000 c
- 6.06
5-4 0-4
as 1 $ .- 40
063.
4)0A 3 0'4 W W 0
66Of0 0 >14' 10 0 tn.0WE: 4 m0% 0 4 &0 - r k- 10-
ON0.r4" >4 0c~f 05j$4 4 1 0 41 050c34-d 14 so4o 6~ >0
In0 41 P 9f g4.-r 0 6110- a 454 0 N.4 4-4 3V 0 050. 00.0
M'-4 -64)6
011.4 w *6G044~6 0 44.4 055
10
4)C566in1-.-V650a 6
0-644-6 64.6504 4 1-406'. 604)4 -4 -U
w)60
4*4 04 46
666.00
o14 0. 6-4 4)60364-Q 00 6 -C 44 . a001
5 c
4) 4 4)4)4)06 ,
0 ,6CL MO - 041 05 .0 0 (0-.4,60-444x.064)
M 034V060%M 00a
E- C -C06
0
60 6.5)6-4
'i )
~ 1
6 4 6
040
-
~ 66
560404 .3-E6066to 6V6
* O 4
0.2
4 A64001.
0 6.4j 06.
A44
N5046- 6W 3
U -0 k ~ 8vjV%418 54 r .0s
0 9 % . )6 4-.--- -4
4 0 o k 1- 0 S - f
o~~
*-.~~ .- a
6
.4 i5*4
4
t*Ii 0 e
Do.0. aq6..f
4949
044 5
0
.454V 1W . P.406- 14;
04 0 % 4 . 3 r -ii- -4 t e-
Sj $40 0 0 3 a
-4 He SA kdj
I ~ 4 I
B- 17
06
0 > 44.640 Q a 1v .6-- C m
U0 0V' 0-14 % 0 C 410 -..- '0.-4
s 0 4 V 00. 16 k
-4$4 N 01
00 40 6 %4 1104 00 06 v 41 l 14 S0
0,~ 0A0 '04. H4 0--44-6 A.4j
*~~ IC1 a 10 40 r-
3. 000lo Ia 4j > 54 418~4C0'-1-0 k 0 04-4 3.04JN 0 IO)0
00>4 040641
k-0 643 00 C'
*.410-.a0C 0044.. 041 A4 L .
u
4- 0 0040C4c 4 *.c 0,5 06 A24-' toU- 01 00 4-4a1 4 U* O0 Z'
z0 I0 aO C 41a 0 0 1 0 0 '0 H(n to 0 4 0 a4. r1.0' 0'
1004. .- I
HC0 ,4
Hg a0-'-- .006 006 - 44 1 0 00 040 1-.40.
0 'ON k 0j 0 1 410C 0 6j01M0 4 0 0 0V.-4P 0-4I2* 7 146 Q.-C 640 C0 0
'-4. Cc0 a-0 J0Ca .- O> 4 96.-oi V - 6W-4 04J 0% c 4ww 6
-0
00 .0 11 0k >.4J0.4 4 , 041C4 ~ U144' V M 000c 143.4 14.400 MOA.
(A 4 140-'a 4 H
46 w 4 0H01.M.%14.-.
H44 14 C 4 003.00 010 4
-41 06 s 0000 0->"43 4 4 40 $
0 4 ' NIkCO0,4 4104to0 .-.- k O.0 0 c 14
4- w10. am C' 0 1E 0%4 r1.-IW ('- 4 O 0 1 >.- V0 a - r4 04,0 logkm r. 000
-V - - 2$ 0 0 14 1 4-
-H 0 . V
O-I004 .C6 0%C 6 01D4 34 0- 0-4 *'14I- NOaNN >c 6 '041.44
NO-40 0 01 0 1 - I % 006 16 > . 1 . -400
V
>00E-. 0014 0-16 C% 'aA 0%041 r14 XH U4J C 0- C ('-.4 010000MOGU 40041
6 610-k40 0-A 0-4 W%0 A1. 44 k4 0 .60 CD060
4J -GO4 $6040 0 014 h 01 5 V00 'A~. 41 . 0t 6- 11-4a00o>0o
k01-.--4 1- M 4 1 c
0 t A000
0,4 c% a .4 -4M0041 0 0 -4 0.060 4 0
.. iIc 6 k*:-400A 041 '0M0 >, -. V
40 .N -4 -1P20 r 40 C 0 m 41
4JCA30r,- CJ 0 00t - IC0 0 - kA 2 r. 6n
a1 1404I-0 04 1 0 0410
0 totatocx 0 0.-.4 000r0 3..0 k 440%C
043 00 0.. -411 -4 0-0000-4 04j Iu
4-na 4006014po 0006a 00.O H04> H60A04101 -4 a
k41..-0> 4k
0 A.- aUO 14 m13rvm410.40W14C6A, 14H -N0.. -44 0.4cs, 1. 0v'
0 4 jH i 4 w40 144160 -: t, a00
'0 4 r001 A$ 4 U
0- O 10 0
210 41 0-4(-' 0 A 0 Wr .- c
0%3.A010 0. 04 .0 1W a-* - n 4j1aOM 0 6me0-4 1
CH JIc '0 -P %'0A C 0-46 a044 0.NPQ ( U~ $4 a
C >N 40. 04 W% TA 0 1 1 C 0I 0Q0 §0 C.4 00 00 NG 00N4J 4 V 6 0 0.
14 0 40 0 AJ
'a-I 20 a MM 1 0 N4010 a'0 04.)'a6 02 11 0 S IO 4 0-G-4W ,a
-4-I0 4 40a 0-4 V
E-24C->- 144. 0; 4 0 3 k-0 r r.0V5 - 4.6 0 400V0X0.. 10 -0 -- I" C -4 341--I
01H4 o0 p0X aa 0 0 x0
0 k.- m4I 6 0 463.40 >
0% 1-1V
SO C4r
Q,.- 2 kV -.0 0-4 0 0 0 0 0 u- '-a c2c - U3C
NO - 0 0C 61 3 4J a0 0 A003k 0it . ra0 H 0. .kWDoc 0W
z W 40- P,400-
N O IC 14 P 0 IO('4.1 >, m .to w 0. In 0 -lN 0 .C0 0 .04
4O CO I0- .1 00 0we
V..-44a
40 W. 0 .40 0%.-1- 44 0 c.A0.c
0t 4 ONOw'0003-4066 c 41V'0 0160IV 00(>('41..IC411 1-a .X04 0-A OCO 01 000
.N4 M6-0 N O-..-O0 0 to' >10 *- 0 1444-.2N--- 0. 4 4 141C 01' C4.00 4 CM141
0 0 -14 N0 -.4I4 0.041.-.4 - an 00Nu4 0 '.
r 4 -A .00 '004 014
044
N01'D
04
0
0
0
14
0o.V6
4 0010
6006
030
NNc -
to~
60
00
to
-W.1
41641
U) -41010
-W
H41S.
t- ) 0 60.14J 4 0466
O.401 6
-40>"
0
-0 C460 01 OQA4; 4 40 001A 14-. C -- 0 6 6 CC. 04 aO Ic 0
0 a.C 44 -4 4 N4a60 a3-0 G04.3 E-la,
CO4, 00 5 'r N>o - 0r E0 c .6 0C10%00 0C
0 qtocV
Wm 4 10>>- 04 t 4A
0 H24 C0040 a-0
14914D0 66-
C0 C0.A. 0,v0 0 14 0> 0 0 0904.4
0 0-I 0
> to* .4%4'0 r' a 49 403.C -00
1 0J4
-I O4
60AO 20CO 4 in 14
V E4('H.4l .- 14
Q 4Q- N40 r0r 06 64 a4 .0 0
0%0 1W
I~ c %4 0 cm - 0 0 -0141
O O-.
C 0
I -2C 0
04114C 1 6 000 0-
6 Mt101 41 *O
06 a
4j C4V14
40
a.0 04
%4 0 '0 0 0 4 10 0> 6 0 '04 r0 46 4
mm 0 014.614 0 a060-400400
V 0 P4 001 0-4 O 41.
O. 00
* 6a, .. % 1U-l43. 0, 6 C 0. a0 a 40)
4 CO a10 to
446
NM
0
~ x
60 '
a00
0 0 4x
c-4
0 S
404.
0 a40
0 0.0
v0V140-.
> 40
N
'0
Nw0000-
CM14
-4414
410%*1
6040 4
r.-0
63
0 3. o4 0' 0 -4 AC0-6404
r%6 N4M 3.
02. a 0 0 .4004 2 041 1 00V 000 w a U
AS w
0. ONJ 0 4 00 en 0 0 0 C40 0 r.0u0 004 a
.0-r 40 0 CL- 4 - V0jo0r G r-4>% w V0
44644 4 . 641
4 0> 0T -4A 0% H4I 1100 .- 1 0 141
0
0
N
-41
0)
a
0140- W
31
0 a03.0>
o4 01 400. 6%4
~0C 60
41 CO am
0.4.
%040
0.44w0w0.-0
0 -46a04 V.4 . .-. 10
00 441 0>.& a 0 . Ap 4J A 0 & 0 04-6 .0 00 444
0% 1410 w9
6 S0 014 oo m 14IO g'4. 0 414 0.- c 0 4 -44a
I
.W4 C414 f r6 04.0g.e 14 me
1,I~.-.
4 W
6.-40A
N 0% COAI'l~~~
14 00 64 6 4 - M -4A . A 40 41G
0 * .010 0 0-414-I-c I t066 40 c a00
-4- >41146 0 I
z40 Z 4 N
00149 t -I04 - 4 o 0-40 0%:.-I 0
0 0 . 0 jVx oaw- . 0 t ).0C00 J W1
6-l44 Tt - 8 -,0 0140 0410 g00in k 014
6 4 a 1- 0 . $4V$060C$4A 0-'l t, S ckI
0 '00
0V 3 0ka
*.W is. 0C0 0- 14010
.14 t- a 40.00 0' 0
-4 NA
V14M W
oo-
A ,
.
r A
-0
0
0~l~ A
"00w
.4
0
~ 41
i. V1t
64. 0
%
4Wo
Sor. N'a4
0
$4! 0
14
0
0 $0
01466 -l
0 H1 .0.
- M00.
4*
"
>H I S 000
$ .P0-4 1 W40 off 44- 6 1 4 -
t8Z - ' ' 0 0 C M 0 0 444-40
0. 0ob w
.4 .0 -. >42 U, r",4
8
"40 % .00100 X9 0 e* "A 0.'0 -00
J,'0004
0 r-1- P, C4I
-4 a .W 4 I.01 410:
W 0 . w & o u..
-- A !. ..
d .. 1 %4.444
4163 a c3
06E~
B- 18
14 -40 -w
4
-4C-. A~-6 MN 4fi 3 4 0 1 6 C
0 ' NH 4u-1 ~de4 03
4
02 ' 4
..$44
3 1 .
,I
n .. . R a CI O
G~3
3=5 Q
..D3.~
4 100
~
9 0 to V 9 0 -
o o4
44,
1z 1 ' a L* 01 -40 0.
ac14 04 014 0 6 t140
404 6''
.414* a6 4 60 0-
0 8 4 3 0C
-- - - -21-4 0043
a- . 45 M a :4
- i -s o1
0 -1ClIA .'- 0
643. a
44to>
4 05 14 4- >'.14
413 10 4 30
Coo04...oo *oa-i 041143
.-4
g o .4 4. .1 0 0 14 0
.4G4 00 O
41
-cW. I %a 4j4.j4..aw s 43 06lZ a 00 64 -4'
00 040 COO0: 1-0
140- .40 AN~
84.0101 04
cm a -4c!;
o:3., tp 0 IS
isC 0
C. a 4 300
00 4 0 -034 a4~ . 4 . v LO.4
CJ.-. . 4 149 X 0
'84 cc . . 0 6 0 4 C 41 0 -.0 0684
1-
>1 1J0 .0 ;
C
mr
0
'
- 40 84o
-0 v
--
"
444
.41 w
~ a- -
r
0.8
1
M 0
~ 144
24 6 414 Or-.-
IS4CS
' -
04_ 430 0. U+-4. 0.4 040 0 4 0J
k00 A4 0
*.~~~~14~ 00
". a z 14 1 4 4 0 - a 444 0 0.
0 >00 00 046 4 6
C6 0 1,' 0'8 4 J 0 i H-.4 0
4
ji~601 w
4006.
-40om
W0 a -
C w~
'30 . 0 S414% -
0, C C .0 CM r. 0WCC *6 *0 3 1
43 464 43*.f43 - CC4 043 CO 0.00
S 0 .0) J: k x 1 4360
to 3-.. >' 641C
to S 0 43 6
143 6 43 IC * 1 04 6- 0, 'a-1
C) 6 C4W 0 04a a 6 0.~ 66k 0. 003A 41 a 0
6-- -4 3C O3 014064 0 -
A0 J 4J'0 C 61 .- ( 0 043-C.
00 14 0 0. 6
-4C C 66 . C430610 C 31
=a t 0W. 46 0 -.C
- 14 0643 4 3- 4j
-. V 6 k403 00 00 3
41'A .0 W C4J
-40 0 S
-$ 0 43 .14;21 a0
4 a M Jw 'a 0 0 114 14
U 6 40 31 -~10,O 00 k OCS &C d
14V8~l4' 0k
in4
0 -0>4U>1034
C 1004
.4 04(U1
>,, '0. C14 1 A . A 0 0 4 - CC
C.6 1Z 0
06 6 I0 64 - a0 0 4-C1Q0 - 0 Cj
3 0 4 w 4 3
6 -I C - 3. 0 8 0 04
140 6 4 .M -6.00
3 13z3.4 0006 40 I3 I 4.11 0 Cix4 3
>. 3 0 . 00606 0
00~>- 66)~
43 -- 0%0 IVV O-4 4 4
04- 1W
0...0 C-1 0 C -ig,4
C. C-0 4 k0 00 W.0404 t" r.-40 0.Ci~ $46a Na q043
0 0- C40 '-40
a40 04 C 80 1 C. *063
80 le 14 NCl
41 C44 64I4 60
46 43 0 C4
j4 6 4. 3 0644-4UNW
(4 Q- 0 C4 S 64W4 '0 r4 4343 04
8
tp84 COD Z 40 -46.4 M43 C 0 0V00
$43 4'4 6 2C 14 04.)144 on C -- 4 -6 ' 43 06%D6..
GO 64344 06Jg 44UC 6-0 4 40A 1-0
6 4.06a C 31 .4 3
6 *0 0-: U) 4 . 43 40. 4440a 0 0.
H 0 c140 0 -- - 'p44 100 12 0. O 0.
%D0 8 0 CO 444640- A
0
043 0.3014 04 '
-'414014 4 164 .-
'- W 60g. Viet%4-44i
000. j~ 4
.1
%z.3 5.6~ 441
I~~~~.
l I. - 1 j. - 4 . 04
;JO
vi 0. t
B-19
4 - 3- r4 V40%04r
>04(00 40.t-.4J4000U0O~ 0 0*-44
44 ) 4
04) M 4CM .30 0 m 14000 04) a
'0 0.4 0.4
141 104 1 440 c 0 400 4J
10 *11 4)~'0 04j 0 00.44 0. 004j0
4 114-0 t 10 c)&O AC0 0
0 .0b. -A cc> Q-
41 gc 040 0 00 C
-1 M
41 c 00 0 0 M4 41
c01400 .41
C 0 00 061 mo 0 41 w ---Q0'V
3 C 00 E
1 00C-4 0 W -A4-M411 C100 Q4 1'
0 0 6.0 'a x 0 a
'0-$4
w 4j1 k 0 0000' 64 4j
~
0 0'.40'0)0. ~ ~ 9 w4.0.60a 0'.~to>.
4
0
A
0
~05 C .
~in4 q) '
1.0$41.
04j 0>C4
M
4.000184
040 0
04-' o0 00
0M 4 0'
4jw
C
1W
4 0 6 4
0
0
H -0.4, 00'.(D G, c014HOOCCC'4
R0 o' 4'j 52
0, 4)000.0
0 4)*.
ICm400>.l 4)4.
ON)C 144)). X 0~o' >. o4
(04)0-N. 0 0 C A 40 0r.> '0440
0 000 440 404.)0 .
-4~~ $4 -4 0, 0 0 4 00 4) r)44 54
.0 H1>.C 0.0
W1. -. W1-,4 *.0.4 4 0'.40" 00 CO
.I)0. 0)0 0 01.A100
r( JJ 0 0 0 00 0.01544104 C. * 0.40 > C 0 V0.
00 41 0.0 A4 H
H u.0'
05 4)0.
.1-;o00
1.4 00k C 0 0O .24 ) 0 01 M 4440) m C6- 4'.
00 0 (04404
z 04.4 044 4
4 t00 0.A4J
40C 1N -40 ri- 0. 0 0tow 1 CO.AQ
m
0CC - 0aNO l
.0 41 0
~~~
4Q 000 C'- 0 000. 0
'A00414.4 m --
4)1 .: 0 0.2- M'0-
0 1414. $.40
4) 14 1 'AC 0.0J30 CH000 41 4.4L '0 0.0 0' 4000 c L '0r
.040 ' *-4O W 0. . 0 0 C 00 4C .--. 9.. 4 0 -4 - )0 U
r-0 0 0 - a0
.0 WO go 10 1 'C00 a1 *'.4 4 (04 co4 0
W4) 04
0
WU
8 W1
C 0' >C
0 0
0M
44
0W!
w. a4
4 4.a z c4c 400
0
00- 0
- 00
04J
44 04C
C4 Q
00 $.44 0r 0 O 0C)
14 r0 0..4~ 0 0 004.u. IV *.00*0 -0.c cC
000 0>0 0
0 414.0
01 4 14 C 0'0 00 0 0 0 44 4J *4>>N C014)
O4O c)(000 .0C rOOC'H 400 5,V.V
-U 0 l400 0000
*l w r 44 00j a, 40C 040 00
r'- 414 AH c00
140~~.
600
44 to 4)00 0 .4 0 W .
C 4)0"0 01 042 r.0). 010
01~
00 .C
1- 0
0..0
m.1~4 c 0 60 & 4014.0 c. .041 1 (O 00 I' 14. 01'
a, rNOON c 0 .4 .06.0)4 m.CO 04-0
a> 304.
= . H cc M4 a-to
2o 444 40 0 4C r.400. Z 0 00 4 >.0' 0C
C0 '0 -4* 140 )> C'
-1 '0 <.0 -04)
0.
00.1w r)1 0 .4
0
; 0 0 a410 4
Cr01 1 0 C
040. P c0" .04co0 0 .4 0 a vto.0 4OC. 4 010001 Cr04)
04-M.0-4 O >. 04;0 4)00 504) 4
X'0 I0 40.4.
- )a 4) 0*-100 toOC 404J 0
00 4)O
w c
-'
.404
0 N4)90
, M4
*
~ ~~ 00
01 Z U
CO
44
4 0.00
.0 0g
0-
00.CC
4C-.0 4
000
)..---0O
r..4
'. w 0; 4.) 00 .0 O>.0- 0 0W40(- M4 W)O ) W'-O440
(00 Ho
0: 00. 14(a-44 O..4
c -4 04C 0)0104
1-04 m 000 C -40
'm 40 -. 00 0 0 41 c00. 0 0 W 00to.>
Z-40 0
w<4. 0 04 0 ) 0 <to4 tCor.4 (a0.0 VNO
41 .40. 0 Vr0 4 L
4-0 14 14(0-4 O.00)H
r 0 4C : a - 00H
0'
c W 0 0 V0 010 C X M 4O.0.4.%
>.. to CA CC0 0 M4C-ca 1400 to)c EC;V40w' 000. 41 0H
0
0 -'00.1
0 6f
40(
0
~ 4
14
0 .4 .
A4)
a 00I 0 0 %44
'. 0 $4 &C.4 0
0
w.4
Ct r- '0 .
04).0
H0) U(
-0
0
00.00 1 .
0, 4)4)M 43do
~
z20 4) 00
t0"'
L14-.Q
H 0414 YO N0 VCCO 0 0 a441 ".0 101 0o 0 -400a .4 '4)
o $4 A.
-00.4001I4- 4w00 0
0-00- 6 go- 0 >.3'e, 010 .0 4 'a I IV6 0
00 -4 0 0. (0 000
6.' SA14to
0 'H 6> )
-4 .0 4J4-A J14J 0 54>.4J a,4~ Im D 0 .0 64 4.0'
ilicI l:4cIIC- fz ij ogjj ! j4;j U) f 4:c7k
w CA01 U 4
111 W 0 4 m .c -4000. to
%4~~
C-4 C-
04
-AN M.0-.5
% 4 4Jy 90 44j 1U0%, 004
ca O $3906 aO..v :0 tyO:0 v 0 = e k4a I
C4 00 MO 0.4 k6 aC 0.04 aO a nMO 00 a00.
5004~~~C. 0 000C.~0O 041 w.-4
IC O-4I.441 M OC cc1r .sgO-
aiw4-4
0. ~~V . c ok10
A~ 44
Ci-3 C A0
w 0 e18 0
1 !.0 SA01%S 14CO
-3 0V0 C 0 0 awM
16 1 4*
40 0V4 0
3.0-
- 8A -0U>1081
C *6Vs
44w N WO ON44
6. 00040 1: C 9 61 60a4
00 1 .
6-144~6 M 04Z 0 % a0C -4
M04J-'
0.- -
-4,04.- C
MU 6
0IM 4.400
tiM. 060 MOO 4)iM>
1 4U -"M -40
. r 4 4 0 2(0 COO '4C 0-1 0 41
V .044P60
a-Oso4 gMt-046 69' 0.00.c
$04
Q U)64 1-' MV 0' 0 0 04 k -43 l
0 CO 41 41CM
a V.0 f 6- M1 1
04 MO 'Oj
0 4 C00
.'4O~~
1
.
0 c ~O ~
-..
. o
M00
0
04-
W0i Mr .
04 0.
4100~
6 i 4 0 OO*
1 6
~~~~~ OC 6 4 4 -o0M ~~
41 0
164 as666..C10
-I~ ~ *61C.. 9.C 4 64OU
M
0
CD -4
06
6
CD 404
C
*
441
W 43 *0
4 OVi
1( .8O 664 -4 gCo0o 0
M % A RVi
to.44
H0-
0
H0 C I0014 .45
4V C
1 0O41 00 M-.40.01
1, a c6 0 w :Vv -40. Li--
O4 ~~ cc0(
60- I
C650li
M O O-.0
V1 61 00
0i~0-U
0m 4 W~
4
0
0
. .440
r0
MC
r; (A 0
oat..
0' 4-0
0M1 0
00l W0
0060WO-a
30
~
10 Q $a1U 0 OC
6
01 4$
0
0.
C
0:
'
,
1 r.ar
VsOX
i 3a-4 04$
0044 *U -4-40
.0 Z011
*j00
4d c0. 4i -4 0 ki V1 0 Uq0 000.
C)
COT
UO 1
-s
0-4
0'16i6
A
6
occeM 014 A.J1~.a)-
40
~
w 4Ma
41 60'L001 41 4. 0 1 06 C 441 0-:. 0
60, 314 60 04 4 C40 A4 -40 0 4 *n 040 p00I
441
0 044~LC
&-4j>.& j w4.40 r 10 aM 0 C-sM-SO * o41 A 0 p6"0 0'.-0
mY z aW 6
A w 4C0>-41 0 4 0 . 0 4j -scO0C e 50 0
V 8 s .- 0
CU)!w .f0 jv COC04 k~~ 1-.
as C0 06 41 M4.4 0 - . "00
$4..- C3 . 6 6044 OV0Cso ko. 040&a0 * 0 4-.5 .0 6
c' :3 14
0 0
r 0 V 4 a >.X6.0 1- -4C 3 oCOg 40 41. ow4414
n 4 t VVCS 0 50CP. 0 0 %''Ci6
0C Mp 0
V0 C 6 M-0 O W *O
4-5-54300M
14 A~C
004
~ w
1C 6
4 6
66.
10k
*0 .. V4~404 0604 C
4~0CCQ'6 0
6 0410440
i k kM
4
00
004j
-A 06 aS 4
00OO 20. ;> 0.40-a.40V4 o
.00 C.-AC.-s NO 0 -SC - CM3.0 0 -0 .60
H kg4 .. w 0N -1 Mj0C4 41COMV 0 14 a H .-- w0 $S46101Cs
.40~--2 0 06 40 %eO C 4 00-0-A. -5 CMM S 0V0 mo10 -4. -'4a
z1C , 00z . 6.- , - 0 -00W10.0
a0 041 a-O 0 a 0.
((O.44 0 4 6 0Rf" a 60.3
S (0.
0 0 4 C- .0 14 k010 .
U0'-'144 4: .%W Li604JM411 4100 O.'-- . 0 - 009XM u a
a iM4 S~fffp44
C~ ZVM .
0ji
4CA
1' ' 50
v
I
10
C
C0)0 0 14J W4J -*C r. ID *00 0 4W CV4rV
W C.C0
041.1 00 0 0 V' 0 44o
1ol~44 1.. 0 0. 03 0. 00 .- 000
4)f v0i>.o-,
aOC 1 044 0 -
u o 01--.I40 41 41 , 00104 04
.1
c4-1C Z-C 0 0 IN 1.0 0 r0 41 0 . 4J4C 0
*0ECO 0o- Wr ( ) 03 DM(D1 : V04$4J
-4 0. 0-4J
4- N .. H 40.- t-A 1404a O .4 r- w 30 [
0
4J W Id , :-
0O0a>0CpI 1U U 00000
.Ca OV"0W1
V 0 1
~ a
OCO 00
C .o 0 a0 O w~I 4.11 0o0
4 Cr 00oo
0C 00-
.A-
toI.
04! oC .M0, 9.0D
00w
r- uC
-10 >~ 0. c
.0
00 (D
0
CUa
W
. 41> 0
rll0J
A41
0) w00).1
>. 0040
)
N10
U r Z W-400
-0
014
v .4to
00W0A
0 to 14 r
4 4J4J 0 U4J 4J 0 4J 0j) >4) 01 -M1"4 124 -C 0 0O~ 0V 1 mo%
-
r-01N.CE.-44.1L COC 00N O f O w~
0 0 4 to - 0 C
:3 OC10 ZIl0 C 04. 04 ,C 0-
C VF 0 1A0 P4
> a:D0 ro l C *0 *4J $4 0o C)>
C. m 1 0 D0 04.1 40 4.1 0C 3p c
- 014 4.. 0)- (4 C 0 V~0 IC 4 4j 0.CC0 4014 00(a
-H .0 r0tn(1 0)1 0 - aC 14.C40 0 0 4C0P 0
$ .100C. 0004.
'0 Hz00 .(Z0'0 :3 0 C4- 4.1041.to 4
z r Ca) 0 .4 IC0
(a .. O * U) : C 0 1 IL M4N :
4>00.) .lo r..
C) 01 t0 0 C Z 414J 4J-42,
V 14
N- 4) Wo.014 0 eq .- C4 0 .-
r 4.1 0. (40
10- 0 C VI4
0 0 0r 0 4
0 N 0 w 4 U 0 C C00(X
2. C.1 0 01 0 r.0
0 1 . to t o s-o C 14to , .4 II
CL t-00-40 to to0 0 00 l0a
x < 0 .l 0 0) <4
0-40 114 00
ca 0 004-I 4.1 ca0 0 'a0
0 co0
-I 01 4 Cg
10 0001.EI
0 0 0
r 0 041 01414 0-l 0043ICw01r-
0~~ 40 0.0 04 a1 04 4
14 co 00144 $4 n to r4 w 0-IV-
-0 04-
0 - (04~-
$.41
(D- l' a3 E,4 4) 53-V- 1 00 - 14
, C o 0uw
$.~~~10COIf 0 4.1 Ngs 0wc
4)c Im. &, ~1 -4 0o 000
4 I 4W4
0 0412'(0
1400.
'0
4 0 A 4J AJC %
IA014.1
0 4,- r I a 0
a, r10 I1w (4 fa.- 5-4 00041. 0% (P.0 C 00
u a
0 ~~ g
0 1 A
140
0A
Cc 0q ty40 q
0 r. v104ON
0104.
0-03
4.1A
>.I
U) U4rIfHa, M in w O4SO .- ON
0 (2 A 0 .CC-I SIn-
0 -- 40v NO.- 414 .4. 010 0000-0
- 0 0 -0l5
-1...-
0 u..000
. cn041~ 4 - % 4,-
ONIA
H w.00.-
414~~~ .-. 4,4 0U- >,0444
B-22
0' 4 -4 n o40 0 % 4 0 9 4 00 0
*.4 00 -4 -l 0 4X4(0
- l
0~ . v4 I.- C 0.
C 4% >%a 00 0 N 0C -044a
0 O 009... 41 .4 C, 4 '0.1 0 0 4
1; 9L 0 W04
0 ,
ON M 40 0
-.
(-4 En Do A 0 0 to 0 @ ..4 0
4) '- 0
~0. '~c-40a 00 -a w
0
Z 04 0.440)40.
0- w)'S a44ta04414
>. 00 lb.4~ jcb 100- M 0 .4)0 041 00v0 60.501
V .c 0.Vror
0 r.
44 XC 4 Cto
44W
44 0 >4 r. U3 r.0 Q 'A 0 $4 ubIn4 4 C$
tl0W.0 40 04 0A'
u 0C
4 41 1 41 .0 *404404) 404
-3 %04 V k
M 6 4-4 4O0-4j 4) a so 0 '1 M 0 4.0 0 M0.10014V 4)O 0 0 P.0
44 0 9 41O O 0 C IV >1 R*'-0
4 V COOUO 04)4)0
0l to '4
kV0 00 to 0 0 )o
93
0' b0 *3 4 00.0 0 4)..
V0 400 0 r 0C 40$440
0 z .0 .444) o .>a 0N 4J 0
044-44000 I
0030 0> I04 04
44~ t0 W" 444404r. 00 0 4 0
U0 0 V4 lb 0. 0 0
02 'lbm-4 w 00 ( 040 0 a)
At00 >0 k4 00- "3 0.4
4 00a I
4004U4
04) Ad 4 0 0 0 00a .a0W0a 00
C440b440 > 0 w0 90M4CU 0 0
4 3 00 ) 0Go 0 0 00 w
M%0 4 ... 0 0 0
0C C 4) 0 0 O40.-q)0 a ?A C 4) 0
0 I C3 0 *- , 0 4 ) 014 - 4C0 44 a 0: 4 0 J. JC 00410 4 0 H O0
04a 0-4V V L
a .U 44 0 04J %%
4)40 .4,0o41 0,4a u03400
000 4 A0.-4 0 O14 -4440C 00A 0P 0 0
4J
4)). 0a 40 04) .- 4)
E0 Q 4 JfMU 0 4 4 00
02 C V 4. 0 G O' I- 0 14 4 40 0. $ 4 04 040 0 0 4400 4 0
000 43 44 00. 00&04 44 4 >b
04.. 000.toC 4_ 0 00004
0 >00 $44 0V.4 0 40O
D'- 00 $4440.0 w 4)1)
00U 01 0 w . 0 0--0. -0.4
020O r. 4' 4A4 C 4)40 11- .0
to-4.
044 00 OA 0-4-0 '0 a 04) C) 04u0.
0. > 0 41) 4.%0C M '=44OR C00 lbA4 0-4
HO -r4) V .00
44.
0 0.0 ".m-4 0C.1400 -4'a00 0 a
a 0 20
o4-0r, 1 14 4 w so 0 0 4-4 M '
0.4 *OM 0 0 0 03.s4) 4 0O. 4.4 -
C4 0
a$40 5 3 -4
m4400
.0
00 0C 0 r. aR4 i 0> 044 04
044 a 0 0 00e U 10, 00(00 0414 > 0 .. 4..
. 0 k M.000 9 .0 41 a > W00- 4141
H 040C 004 046,0.0 44A to; p% 0 004VC40 0 0 4
' 03 c 04 $4V V0 r6M0 - FA 044I 040
M44. 0. 414 w0444 4 00 .44 r4 0-4 04)
)-4 0 a 1 -4 04 4 4 09 '000 0 MO)--40.40
w0 %(04 0 a4)40-
H
-4)0 444) 40 0 so 00
0 4 0 0 W1 0 0 0Cq3..-41 4)04) 0 0V 4 0
0 0 4 a 2 w 0. 0 w 0 E . 4 0 4) :5
'n*0 .. 0 0- 1" 04) 0 3.
C>0. (-. ) lb4 4.0041 0 P004)
-41-
Q1 40 .004 20"a 1 0 -. 400 043
00,1M 04M 4 4 0 4)t0a ,1
0 - W 0 10 .' W-b.-4 444 0. 4J 5
4 U00 is 0 04 01,20 r> % . k -440 0-
(n4 1- 10'
0444
044.
0) 0 OA44 jb0- 0 4 4 W 0-C 0 , 0;j '0 1 4 I 0 'W)00
Oi J - So 0 044 0 -0 m- 0 a.
00.4. 0 V) a 00 4) 4)
0400a
00
P) 0
U
00
- 0 -4 4 00-43
00 04.0
k4f
-a- 04 g 04el0 a.
00-4
U..044l4.4f 290.-
. 0
c444@-IM
a
0..A u -
0
0 ,no a4' C a
0 0 0r
-4000
go 44 .Qr
M'.3
0040
*4 Z
4-40
4
4
0
-400
.
4
40
o ~ 'A 000
4"5 o% - 0 00-0..
4.
0
0 u . 0 .82
04 -
0M 4 -o n 04W4 04W)O4J50
V -M0
'4a 0
u0' 044 o 40 3. 90 %49. 0 40 0 4
a- k00 0-. 40%
0O.41.0 1 41jj0
14
00 106 4 0 .4 *a
-W m4
00 0 4I 04)V0 04Q
to0 04 -
4A) 044
0 ."4A
>0.
O -1.4
Q 1"
In4 00)4 .b 4 00 9I -
(n.- a 0. 4
U000aM1
4j c)V04)
%4 Mf~4.b
4 0.4 1
a0441~W~
.4 0 R
B-23
a)- U) >,a)t( C0
r"40 Ow
.4C W- g 9 w 0Wm WOO >.C
42~ > XC54 10.0) -HV.)
-NI-I--
m 35410 -100--
a) S44 4 oa C
a)
0 4
0' W
0 15 .c 0 V r0 0W> c:co 0.19
u
NO 5.4 )- >.C -~W to31
La-- W 0 a k> 04)1 4 ME0
m0IT to
041$. w- CO Coc
*a.C 0.44 ccCI . 440r.(
0L
0 04 04
*,c'a 41 r.0cg-4C $> 4)5in)
4) al V0 .4--'C4 > V C-C Wa)
0 c V' 440) *C3'- 4.3 0' -4
a) 0*44 40 4' C .- %.. -.45 C
,,. _4~a OOVV40 44..-al 04) 0 0c 0 0
Q cm41 0 9 3Ct 0 -4 0 -- 0 Cw W
4) 0 41 w
0-a 4 0'W1 01' 1 .C W-14
4 0 S.C
4)4p . w 0W r- a)W -44 44 4)
to20 54 0
$4H0 -4@-.(a0. Z, - 4) 1 0 C) W ".
0 N-S514 ,
114 -4 La 04
go OV00 AoC
W-4'-
WW *4)
A. 0Wa0 %-.C * COO0
0 :1' n aON 0- O 4) La V UOW
WO 4)V1 5. V X0.aW O04 0 U
0>1. C $4
Laz)' ()5Wa 00 o4 a 04m 0 4)V
Lw (y,4~-
c V -s; s01 0 L0. - .44 2WCO ' VC-
tow2~ )V0.'-0 00.541 "1C 0 120 ) W o-4A a) C 0 V c'
0 a. 4)m4a- L)a-.0 C. 0 -43 0. - C0 OA
a. wC60. V) W
WV V L 00 05.- -50w
&C 004)00 ow~
54LCH a5 00
W5 CONO 3v
41- rW-W.- C1 OW
0.3> 0114 0 0% 00. -'40 C)
U*- -4 S.04)N0
05.N
0a 5 -4
4C4J kr. -- OC-I- 00 4J54V4 w ) CD0 00
.- c .N w -M 0049 " $r 41- 00' c, la W
-4C 00w 40a 0' -A 0) 054 WcI
-'O4au 0-s v I C.1 4) -0&* C 4) 1,-
Ic0 9).0 41 r4c4 a .0 (S %
00. 4)0 OC 000 q w 4j - q 0544
-0 00
-- W 0 0410 i -. 11, 0.- g
2i V) to go
4)) 5 Wu s 1 -5-)0 @2
N 0 02
14 05 . .4 atW )0 0 M 4o 54 Ix
m tW
P4 r
k'@ w 0 -*141aw
00 -C>W. ,
z~~~a
4 IM0 V4 km
a' W 4
al ( q 0 0 aW l 0 0-W,
Ln0 Q 0 m 4 kfm ( A 0 4 - 11
B-24
. 000 C
0C W~ 0
>,Z4 o0 A0a
H .0C$0
V 4 O-4 0 0 .0 O40 MI
M0 0 44000 0 0V8
-4. 0N X 400.
C 4 6 0 4!
(A $4I044
* 0
'-> 0 , 01 c 4r Q-4 $0.-M a
0 0.0c 0 p ma04> XE 0 o
.4)0
0 im 0 01.-0 V C M 00.H4
bi- ii CR Eno 0 Sto ,A 04.0 *606
0 04 02 9.4 0 0 H H. 0 t"0 -v 1 '
0 - 04) -40 V
* 0 -.0 .~~0 X.0 04 > .
H0.
H
IN0.4V
$0 04
0a
0 0 0.
NN 0>0 00'-o. CO.
0'. 9 E 0 r.00 C> .M 0.440a
000 0...0.r"0 0 4 H"
0w 0..0)0.-,40
.0, V0 0400ON u 0
0 r.-.- 1 W$0 W >' tE.. 94,40_4 0.
00 C0 -4W 434. 0%40 0 ..-
0.. 04 r. v CL
HC C 000.,.4 r. $4 N .o 0 a
a. 820 110.C40~ .4V 0 rz
>4 04
r400 tr Ho 0. v 0 . 4 0 C a
00 W 0 094 aC o.. -42 >0wCD 4$0 4.0
0 1 04 0. D $4
r4 z
CJr c 0 40- 00W: 0 04C4 .0
OC .400U 34) o >4.0 WC -4 M
=40S0 0 0) 44 0 'A W>
0C- C z.00
I0.ON0
C M
0 0
C0 r
0
~ IN$
(43
H-4
HO.
3-C
>4
0>
.,
C
v 1' 00
4 0)'4 to 10
VCCOM N
50 .. 4 00 3$r 400 :00 C 0k 0 v u IN ..
a, U00 0$4 c- 04 .0 0-4 0 N~4
0,0 0 X -41 W. o Q. 0.o
WCV 43 0) M N r. 4)
T 4D0 I J 0 0 a 0, W0-4 k a 4
4) to 4 M0~ 0.. >CJIN40 O >>V 00I0..0
lt
Z0 04 E-. C N 0. ..0 0. o' IN 90 to
I
0 C 4) 4 40 $0 W0C- o4 0.S z 0a ' 0 4M'
a, 0 o) 0)0 0440 0 0a) % 0 40N Ma44
go. =
,C 440 0 QV OH4> -. 0 . " $ 0.~-0 ( I
14
40) E
0 2>
4
11O$N $4
V0 :> 00
W00D --. 4)
1 41
C -
0 ~4)- 0
0
0
42 a
$4
0>
9
N
>..HO 0 r000a .0
. i> 0. r . 0 U D0.
4 0*--noo
0
04- 0 Q INN-
0$ 0W -.$401
:5544*0f4 0.
0~ ... I
r= 4)
zj
"
1
-
A
V
0 -4H>
.40V0.0
0.0
. $4 0
4, HO
I ~ 0 04 now
-4 (44. -4 4 0 04 A
t0000 04 4)
~0 CMO-44- i4)
z aC C44$00.
0 . aC
E4 M10.000-4.)0
0 o U4.IN
000>.ooot o
0 00> 0,
000l<. 00
.-
0
M r.
E-)4
.04V
0M
$4V00
0- :3
.4
C 000 0r.
. 00.000 . -43Q44 M A0 540
0 A0m4e 0u 9 4. - 4% 4N -0
0'- 0 4 to 3 16 4
.-.
4 V- 0 4.) o4-
C0 A~00
$40 41
U o4- V41- 1.4-4-0 0
0C--00 0 1. 0 V -40I
V-4C$
04 C.
$44 0IN. 014 -020 r..- 0. 0 0_ $
-4 0 r 13)40:3 .
00 0 0, M0 0 ,0000 V0 OW$4 "- V 0
m4. 00M 0. x41
5 C0
u N0 .0 0 0.4A0 $0 w.o"404 I0
k 0 4 04 - - D04 N4 -44 $ 31 0-0. 0, .
00 W4-.
M0"0
90 CD0.
0. 0 0 2 0 0.40 0 - -s$C o0NP
>,4"v$-- . 0
0' 2 >'4.)r 0n 00 .. a004 ..M4.0 00 0 0
9 U va C40
A v 00 0 tc $4 04) $4 :4 004) IN
4)0 V 000 W0.r00000in. 014V)-.) 001
$4 00V 0 00J4.044
A 4 014 .M~ 0 0 41 C .C
0~ x 0
C0o 0 = W V.0 0 t4) .I
.-. $4 V 0 0 44 0 l< in 0$4a,.--- 0 0)0V -04 <4.004 4. IN
0 0~5
.4
w
0.0' 00-s o > 9
440000.-)
H0 0 :31 00.
I0 0 as4
0 M
-44 (N.
%D104
$4L
.0 40. 0
-40.
00
$
00
. ~ 04C
000 0...)
00 0 00040
rO .0. 044 to4 o00 14.
C $40 0 a' 0 H4 0 3 44 -It 400 $04 t; o - 00.0.
C40.4 000 9:~0 V
'- CO0 WA4440.
O a0 W0 t 4 4 0.4 -0
0.J a 00.& r 0 04014o0 .0t41 0 Ca00 41
4 0 9V$4q0 a V .
>..*.4~. 0 0 0.0 v4'-14 00N 0 I V4A4.) O 0. 0 x 2.1.a
0 a 41oC 1 H"4 1 $40 r. 4 42tom
5N 0 40 044a0.)
00 o A0.00000 0. 4 04 0 0 W i . 4
0u.1 0vv.4
S..0 -40 x.04 001 00 4 HO IC MIS
$4 40 0$4
00 02M - U 0 0 (-- IN40
41 0 -. W .~-4
CC 00 t0$r-4W0
Ma$ ow 0
N0 cH E4..4-
00
I C
0>4
0%4)0. 0 to
4-....
u0
I Xtp
C0.0 0 c
04'4
I 40 Ai M
(4 kV
-
0 0
0 4440 C1..
0401
o.040
0 IN0
0 CID
" go$
W
004a
a-)0-10
0,
N~0
r.IN
00 C
"4.00
4-0
V0
"a 00.
4)
V a0M
o0
~ 4
-fu O CO
10 04>,
4 0-4 00 040
% 0-
00I
14 0C
-
C4:3.
. 0, 00
a' r"4
4. o r 40' 00 0 UC.-.a --
4)aI
0o0 H 14
k!..
C M~P 44 .4
-40 05MW
%.444V.UM)-04N4 40 cI
%0 V 0 000. 1a4_
.)o 4 94.aa-) 0
$4 i 0 4 )
00
141_
Q. 004P 4 >t~l r41
-.
4 0.4H P4 WVe 0 02 V4 0 41 W QC-40
0
021
0.0A! A4 1
k4 - V 2~ 0 0 &~
V 00 40
40 10444114)44
N1 %0Wo0 *0)
1.1 a 14i
-. 0
1 00 0 N 643J~ J4 044 0 >, -0 D04 44.00 0 V MO. a a4 14-
0,0 4 901 0
02 U5A040 .wW V I ia > -4 H 1M 1V~ 1-0
u0 l 44
0 V
441W4C 0 4 $ r M-0 40~
14
0.0 1000 0 -( 0 as C C01 0
'-04 00 aml" ON MAN 00 0 49 1 r.
100.- xx 5 O f4444)41. 0 . 0
- 9 l P
b0 A$4. a a. 0*A> -.4 0 .8 o4 0 A MID- a 0O4-
%4t00W C 0 40e0
0, N0~-~~~
U)- 0 600 1444 214 0'"H-1
"0 0
ZT
30a
14
V
0
r4---.
S. U01 0I
S. 0 I zN4 0 00081$0.0 WOH-~~f 0 m -Ao s ok
0
C 42.5 Z0 a V COO- 4; Q IN tP A 4 ()1e 4a Mv bi 0 1 0 -1. .4 41.4
1 4 01 1
0 i C O N
gA 40 v en40 "0 00-
0000 ~ '-4
V
40 rIN4-CN 0 v~'
a0 0 8
1 5q 0-4 9X*4~ 4$1>0000
or0140 &0
0 "4
0 a
0 C .0440
>.&~ 000
-4
VC - 1
u. i 0 A m 410k h-. 0- ) -V- m
0. 00 v -. 0-4O0 .0
B-26
o~ ~ 4i14
a' r4 0.g0.
S4
~ 06SO.U4ou
.4 .0
* ~.4V14W. 0 OS
* 3 a) a.
* -. -- ~ o AJ-4 c4i r
o,*. 41
I ~ S4J ot 0 =4j
HA 4 .0
-4 'aa *U,4 -4.404
.e0
0 1.4 04 .00 -4 -4U4r 0 1"
0
oe .-0
$4.4
$4
1.uo41 a 6
* - 0
V.
.0,0
U) 4 ~ ' 0
.4 MIX. coo-oomoo4
4
- Aj
$4
10
k C.
3 a 40 0
410o .4 0
c-4C0
014r
~OZI04 4U *40 *0
41 ~ ~0 040"' r % 0
c 4U 0.1
49 IV .0 -1 0 a -4s.
a
C'% cn 41.1% - 41;410
N ).0- 4 -Z 1600a
00. t* 4'oN .0 'o00'0 $.4
-4 C01.4j C, 0, =
j.0 .oamo
O r- 1014
a *amna .
10M
10- a
04a 9 0 41
$4.UaU.144.-
to 0. 0 a0V$ ,
0u. 5 60 4'0. ~
M 4 11 41 4 0
41 W 0 11 a 0J-46 r. 01-'
0A 0
.1 00- 41 a
Of* 0~ g0- 0 d J " 0
to _4x 14 0 14
r- 04
01 2c0 014' S4.O 14 r4300 -.
a. >u 0 o a 0
to U4 C.0
0 0 aCO$4
Mta C4 Q e' V
at.4 03
.4 4r 00 14W .
00. 0
(Aa% .0 a
0, 0 N0' '1
a 4 'MO X C rCU J14 0a- 1
OD.4 ->0Na0. 044.4
MO 4 80 0 4 0
0 t 0 )r C. -'o.4044
.a.400 3 c p.
4' 0~. ,
-4~~~~~~0
I c
04.
'4
4.U
0a.0-
0 40m 4f
-0 0 $-4 4 )
144~~~ 2,2S
inr4a
4 4- 10 ueV4 c
C 4)004c004-.
-4 0 H '.4 ag 00 0 0-4 A
IS "I a 1 4)0U4a) 04-' 0
4400
0. 44
a *.-I>0, 04 0 a. 0 4 40 40>
o C004 V 0 V *-4 3.0. 0. 0
44 0 v. 004.0
4)0
4
Coo0
> M0 0
W>,"- ~ .. A
1.,
m 00a .0 A C
o4 0 00
O N 0 ( .4 01
c 4 0 -1~.40 V4 O
0 0 00 4 4
00
41- lo*% 4 0 .1 4 .4
>4
t(2 0- . A 0 V >> 00
41 1 M5- M9 >
.1-4 44 0..0001
CJ i a~ $O 4 C m-.4 4 004 , .'
a C0.
0R44 00 4
-404
40
00 A
*4
(n no
0a 4 0
f
O04 .04 -f 0.0MUO0 WHO 0 w 4 0
r.4
14 0-.4)-0 C
0- 0 POC)
Q
000 4CL#140 0'0 Z04w.0
*'44 0
H> 44 0 c0.4
a.4 v)))04
04
44 go000 r.
0124 44 00.
4 0% 040 010U -24 w l a4 0 0 10
a , l t 0.40C 0 a-.
44 0 044 4 4c 0 . N
n V) %
OC 1 0-P ., 2.
to r-
0 c4 0- 0 4 00.> 0 (a 4 C40 0
41 0 0 0' 0 004)
0w 12IC 4000c -4400tLWC '300 0 444..4
4 04J
o 4414 1 V c
44C k M
10-4H OmC Is(o'4N0 4b.o 0440 0 .4
14 4-'& 9
404 w
IC Q 4
~0 4)c 0 0 ,004
a O006V0 -k
('0 C
-0
.
VU 4J "-
0 to cV
t
W40
a
-4Cy-f4). nIz .4 44 00 %.0 40m 0
0400 -(Oc 1~ c 00 C 8 40
14..' 44-0 0 0 *.C 0 4j* 44V0 4 '0
.4k- V1 A 40 .4 k 4 4--
041 0u 410 $4IOO a~ = r
kN
0 1041 0 $4 ( - 0 4 4) 3.0
00(0 00 410 0 4 *40-4k 400 44
.0%4*4.3 40. 004.. 4) 404 0
V) 41r0
U0 Ua - 0 Ic4) 0.
00
.Z0UX
in(
4) 0)0 awda JJ % 0-
m$4-
.
0~
000.0
0 '
'1 0 0.
U0~.' -0. 1-A30M
, 4~~l
01 00.m.0-E..A,-4 so
4
". C ?
U04
4)
0
CC
.4441 . C 04
84.
.C OI
0 a4 4 00.04
c.. 0 (- - 0 >-4 -4X
w
-4.4 6t w 0 M0 U ' 0 0 4
M 0 go V
4-1 0 40 1
0 w00-4 . U k 9 9 m40 0 ct 44 a-
> 0 A:2.44) .C c
- 4 c , 0 H a 944-
C4 . 0 .CA 0 c 0 -40 44,
0 0 0 a 04 3 4J44 .4,44 £1 C 0 0 0 4 )00.4r03 0
I0. 0 00.0
Q.-.- 4V 0 6 4 -0OV'-4 t 4 Q M -C 04 0
r-0 0 . 0J-0 >040 0-
A-4. C . 4. k 004.W04V
OOfO 04)44 000 -0 k Qk 0 Wl 0-.4 0 00 4 40 044
r. .0 ck .C 44400'a04 0. 0 0 vI .ICO''44a 04404044.w
z- V 044. >'4
-. A2 V 4 00 0 4J4.0 0 C
0. C4 4J 00 44
0""4 M4 CO 44r 0 4 0 4) . m
44.- 000v4C0- w
10 0 >0V30C 4)0'0 C 4A 04C' >,J 0V 0))0
0 0"'
0-44 11. a 0 0c acM 00.00
(2.. (44 0 k4 4QC C.-
K4- 00 0 t4V 00 -40>.
t C4A00XMw v k % -4C
0 a Oa
04
0C
-
0Ck0'
.0 0-4
a a
6
0
' 0 0
go
4 a4
MGM040
acm
2..
I44 01$ 44a
HC
.
0 04444 1
0048;14I104)
-
o'0
0444)4k
)40
Ma Q%4o0%4
4 aC u 40k
£
V0 0 c0.4 7 0-ItS 400 400
U00-44)M404 0k
00 0 r4 0J 0 J40V 04C 0 60~04>. w 4) 04 4J-40a
r444
%4
00448.44U
a4 0 O .4 94 C)4 0 4)4) 4
UN) 0 0" %4' 044V0 MO 4A C UI a c a
%0404. 17 >
0C VV4 0iC 0 0U4 4).4k
4 0,, 00> 0 005.40 00c0 to0
C 0'-4CIO .4a 1 9 k44 V"0 Ma - 4.0 aMk CO
0.. 0 C00 .-00 14.44>)
044 04040
-CHc 0 0 a Ct 00' a 40.4V 0 '1 0 1 0 4 r-
04 aN£44K
04C.. kCVO4)04
4a
'0,-.40(00P2 0 -0 k 0 .04300~ 4) 00
.44H
0 a 1 02k>a9rV4m -4 1 AO !V
H'~ 0 c ~o
0i-Oa.44
00 a k
M44 04M 0
-. 040
g .
4
0
44> 00C- 4
c 04 0414 000 4
E-
'A.1.4
0
S. 41 ~
a0 4 C4
to4)00lgo 04
*)4
i04
0.44
*. '
.
o. 0 .- 1 0 0 3.
I2
.4 0443 311 4 cO544 4 go
'0~~ .4*40 4)0h~
IC 449 . AV A -0I!~~
lh.0 4I
B-28
t 0 Z1 0 Z&' 2 N0 1
0-
re 0 54 00.0 a C 04)0
.,N.Co 'k'o 8 0 00 08 00 v 00-.
0~ 0 c
05.0' 0
000 Qc I's. :2,0 0 . 54
CKW-Ic ct. ~ .0 9 :i 0 *8
00 0 0 -4' 00 .w 5o.- 0 J0'
0 0
a'o1
c45v
41
M
a' 4 c.0 0
4)
'0Z..km
.e o r44o 41
0k
V
w N"
'0-
w'
00u a40
%,a
5 1.
0000I0v4Z
CU
:
0002004)R
010141% 4J 42 0 9~ 0-
0a. 000
008 '00. O 0
H ~~
00.
0~~~~~~ 00 0
0-
004
00
ov.4.
~0.-eH0
844
0
44)0
000
.,4
.. 0
kO.40 to0
4
04.40
0040 .0
a.40)5'004)0
4
0
04 0~4
54*~
80
.0
054
S a~
54
kO4
00 c
-400 %8-4
'N
00''0)..
00 a0V4
54 0-
W- Woo.. k0
0..
0- ).
.0A).0
.140k
o o
0'
4) m 54 '00.k
0 )0
~0 01
5.
C.. 4
5 00 0 0
o
0
00M~
00 0
00-8 ~0> zo
04,
0X0
48.4
0 0400
1 0-'
r 0%
W440
NO ~
S
4)c.
0-54 0
~-4k).
.
$4
00400 0
-0~-o
0 0'400=- a "00 08
%40_~
0
-4
k.0O0 0 .
r
540W
&r0
>000000
000tO
0.k0
0'-4 0 aV
a01)0I4054"0
t1 0
10A 0 W0'a0r
a
40
'00V0c
00.0e
0- .4~
28-40~ 544%44'
.4, -'N-45 I
30
>
003. 40
0I W04.
~ 00
41
00. 0
I4.-
0.
'
owo5cwos
0-4
00) 00.4 0.0 60>. a0F4-48 0 to00 *- 0.0 &10 00.00 0. 0
o 10 a to r-4 0540 o 050 54
8-. 0' ~ ~~~ H4
0 0
42 c)
54
0 0 5
1 k areo 0> 0 41
-0 a.4 ~' 00. 0.
00 48 44i U010 005 .- 4- C kIt 0.0... 0 0
C 0 14 a0,40k400 g48--
0~~~~~ 0 v a
0
0t
~
06 40
0.40 044000-4a 0, ~ vo 24
0 4 0V 0400
.QCO
)41C 0. *0-. 5 004 00 k0
0 Il. '. c0 *0.. 041 'A 4 -4 03
o o o o oo-., o0~ a W 10 a. 0 00 we 00 0 00
0 k k a 00 3'0'S00 540. to)- 6.C 900
4 000..08-. o-0
3
to
4 0.
80c 4)
O 00 00 9 "0 J M5 0.4 C- *00. 00 a
0, 0.0%.4 4 in
4)
'40 000>. 1 0 81.,0 l.0
-
I a
0
0
0 4 0
~ C4 0
k80
>to000 -
-000 v0 g 0
r0.
0 0 V o ij
aO00' 8C a~ a'-0-a 05a V 0 00540 Q, k w4) w a 8. 0 *.4 0% 0. 0 c to4) do V 0
00.
41
048H
e4)--
C 064
0
00040
-. V 0.4 .004
4
0
0 0 ~ 40Nc0
04 >
00840go
00a
99~.
0
- ta
4 - 4)
in0 0
) k'.0,
0 00
.4 00-%4
a000
k 0 0 4)2 41 0.W0 '-0-4 .0
0.04 0 0100 00 0'.-4 5 .s .0O (1448 -- 0 40 0005k 4
0'
'40 02~ 05-'0.000 m-Z
'4~ ~~~~~
-4) -' - -4.0'0 4 0'".00
0 F 40) .010 0
0 44J4-000 E-r o 4'J
0 .n0 c05
AM
g 4 4 0 >,a r.0 04 u0.%. 0W04)4) M 00..-I RC
a 3- 4 V V0.0-to 0A oo054 -0.50 4) 1 6 _g 00zQ ->-4
0M 14 00to 0 0Ik3 0440 Q. 0 Xa 3
zC04~ 00 . )0 0-000 06r 0% aAZ 0 a 0 r C0-4 W 0
c 0 4 1 >.00 540018In4-0
o kJ 19 0 "... 0 0 48 0 0 %4 3 0
2 Is0540. 001
.- C)) -00 054-454 . %
80 =-0 - 4 'X 0 4 03-4J-'00 3 0 0.4 .
0 44 -0 4 0 4.) w0 4 j 0 0-0 4 044 N V.0 m0am T ' '
0003
C
'0
44. m m V 0000 q04 ~ - V.'.545
a4- "Us N5t4)JCM
4041
-
04
0>8
04>.
.04 o
a0C 2
01
a H .0 r. 00048£0 go 00% X - 0 a M 0e405
-0.- 00 0.0 goo0 '00c' 0M 4
0%. 44 0 1 E004A003V
0. 0 o 0 0tr4)0W
1.0
a. 00 0 0 -A
0L540 0 0 Nt 054
0 o a0-. 0 04 U E0-4to .0 , 84a 0 t" 0I 0 Z
.48.,2cc V V
-0
*411
0 V 0 aC. 2 Z. 0'05400
-4 00 0 0 4Vk 0 00 . 404 0 50 0 4J 0 0 0 4
-'4c 0 01 6-0-0 0 k.4 0 -4 S4
4 0 03 0 *00
00004 Z 0 0
0 M.--*.
-4443
C 140 cM0 0c - O 48&
GN 2 H . w 00 .0 -40 'a4 ,0 0 goa54 m 448 0 .0.
0 .r.k0 Ole$40003 . & 0 k 0 0
40-'~~ -.. )4)4 054 So 0 k- IS04>
-Z 4 004
.0 = 0U0V 4 . r84 -40 k.4
a to -4 0(0 0 04 0 k.0 0 654* c4 oft 0
000 c0011 .30 0 0 54 55.8
0 0 4 .44
,Z- 0 2~4
> 54k 80-.
k.CU40to 0. o~ .V' -
(v0. Q VV.X -o( 0 0.
..u: 4 V4 40.4 0 4 -' 9 4 4( c 4H A5
0.05
O4H0 05a4 I 0 >3P6 41- v0 .0
0.' 04 ~ Mo .0M
00 1- wo4j 0 n 0.4~.
1 a4 C40 C 80
14 X (n 4j 2C 054. a a~ 514 >"i 0- w
fu-Cc 00 45J0! 2 3' 08 0 0 000 4854"
-4
0-.494"% 0 so O- c a Nf.0 r44
8! 0 a'
p .~ -4 'a40~.~ 55
4u0 V m 0 ma
so4
48)
0 ~-4
V o
4
r k
0
54 04
U)
. Q Wa Ca
44M .. 04.
S.
ou..axj o4J-.0
a4
aH1
aae C1 M404U4.
a
.MaM40~4
aaaaa,
t -4COA-
00 MM 00:
4 10C .C X 44 W $4 4) to 8 Cto 0Cc)j 1 0
as~~ oro ~.-0 ao S.a 0 - c 4WC M a
141 j0 0a4 -"-0
0
V
No0 A ~ 4Wt& -
0
a4
'a14 V~ k.4
>..10 C0
Ma A4 5 C- 0 MWUA 04 a 44 92 4 0 4 0 4 Ma-0
o a 4,4-.. 0 t.C %$ 0M . %41 a C1:
00
Crr
O.- m 0> 4C 140,aM V
-a
MOCW0
44 .0. '1 C 4
Ct
aa4 M
. M04Ca4a
1mo.. ~ "4
44-
MO000- 02a
'DU 14 9 C 04Jg1 C. 4) %Oaa a M 41 4 40 e 100 1
4)c C %.4.4V 0 ACl .C COa a'4 V 01. 0 %4CSa
aM4E
41 .. o 140 140m a4 N ba.4 ap VOC 41 WC
0a.4C 0a
Z.. '* 0-) 2a I>.a o- -C 04 OM r.o a- 01 aO44) 4C4
.)a 0 I..
Z. 0'.4)
a V to
0 aa 0 a..-
C 01 c4)o4 . ~ us. I00 was4 '0 to -0 a -a
ai1C01 013
o~~t a -4.- tr
a14..
0 1 1. N 4) > 0 414a C44 E414
04 NJ 44 0 4 a
.01l c>1I >.0C 011 lU- 9". go >.U 0 C& 4 .
0' 0: aC 4)004 N a C Q -4V 00.U) a0 64-.c a
ozic C-00 0 go T4M0 V0 0r0o4O a 0W-Ia
01~~~~~0
. - 40Us0 14 04 c0 V V
V)4 >W1 r. 00 00 0V 4 0 M 0o a 0-14 V014.Of. g.1440) C
-C
a 4 =- x 0 j 4) to-41
'o 0 a
W-
0
O j
~ -- aC-
A A 2 0 0 or
0v - E
% WIM->I0.0
S 4a * A%04 4
M 0 0
0 -4 0,4
0 4MI.N4 .> i ' 4
-C
a 0 V
0. > v-Z U .40 1 0 W0 o 0 to 4 -40)0Z W000 a ao
0 0 r O 1 0a 14 0 M 0 0 ,C4 aV1 4 0 4
k-a 4) -4) c 0 00a% 1 a 4J4. 9M> W ,4)0
-. 000.
40
Cjr E '-0X400C.1 4)V v 00 c o a MO4u 4-444V0M1
u za - rj00C1,M 1c04 0C CM4Jj C %.4W - -1.-4 V 0 0 M
.0 0 L. 0W U6N0 0 -4 0 V4)W 0>0 to440 0314J 00
< 0A. 2 do 341 4)4 Z14.1 u. 10 0 1-CV 4a*4. a 4 ) Ha j
C. - .0 v.- c o o wea c0 a4014 um %4a
0.lb 0 r. 0aaa>V0aC
0'-
01z . 0 - -. C4410 t 00 .U)jz
4) ,0-. 0. 4 0 4 400 r.0CZ->. W
40 ->0S
Wo 6to0 0>r- l 00.0 v-v-
100 44.- Ca1 -I4.a0'-414 -44J0- 0 6)A
W -C40 04)41' 5I 0-''CUA .M2 "I9AOM0.VM.C . U 0 0 XC.63
0 1.. 0 0 0m >tt04
C $ $ -CN 14 0.0 0 09 4)0 0 2 M 4 M
0044t
4)0 a3 - 0 30- i %v0 000I - l 0L a 001 E44 040 1 14 0 WC<
* 4. 414;a
40
12
a
0 C40
oo(4 44
0Oa4M4
*01
..
jV)0
3 -4430 V)
4
a4 J= 0C41-4 02 0)) C k0aa
0 0..4
.00-
00
14140
0. caC04--4 10044>0 -aa
30 %
4 %0~4
040.a Q
OOC
MV
COM
0.14 C z0 a- OW MO 0 E 0, 0
an 0 OH &10 .0aa 5 -0 E 4n 0 0 04 q4 4aV0
0
4.0'
0- T..0 0.. . rU)a OCC
$4.01.WCE- 44 $ 4
4 0Na a) 0.4-404040
901 4a0 t" 44, 0
0- 14.4 U0 c -) so
aN a,~Q t
>C001 L) ~~ ~ ~ 01 014 V0- 4 0 0W
CW-o.4 a 4) 0'
(n 904 N4 0-4 O 41 1
U
C'4 CE 0 43 to4W 0 -414 Oa $4 4
0a*4
-a@ 4-4Jaa4 0 £04
W0 0 C4)%3U 0
NO-z C.00
0 144 44 Na>
0'0
>
00go
0-
1.4
~ ~
CS wo4
40 r-$.u0$4>.E.1
a 3j 0%-..0-I
a CM
0WV1 4j 04)3
Wa>0
s 1 0>v 0 -04 04 0
014tV4 U) C 04104 .. 44 '0 .4 *0 0 0~1a
tO a,;,
60 a >14C.kC0
,4 C 00 M.
Vo 00 00 0 a0a~t a *04) 9
0
00
-C
W 0-E
0j.0-V)Ic a 0 94
0 2 4IO.14 aI0S
-4
0. 0'C 04 H 24)
0O 0 00W
4
-A
k0 4040C
0OIII.
450-0
U0
140V0
9 V.Q ~
A; W~ 44 W "0- U
0a -'4 4- a' 0 M ' $0a-0
a W jaa 4 r 0 0 u -P r 1a a .
N4
4 0.0V 0 014146 ,CI 4"4 4M *4>40
2. 4 q0. 4
Q 40 W-0 0 041 .4r0- 0 . -A--V O 02 -4
4 4)
0 c N 3-.
v01000C .
1
14a4
~40
0-
44
4)04
a
O 44 0
0 6hr
00-4
3.W..4
%0 _V 09
14f% 00
CDQ3
1 r 4 n014 a$ 4-0 31 C .40.4
B-30
.4
>.4 z14;00040
06>a.rrAw
& N E O 0 0 0 44 4J.O 0 0 :
0 0cE 430146C
0 c 0 04J44t4 00.444 l o 0 'A W
.4
60
4.V0
1414 H 0
0
~ ~0V6.4
~ 60
$
004-4
H9
0
M;3 0
*244
0.40 k'
61
g
O$ 0
c4 4j
0V.0,.4
U6 0 N4
0 00 0
0 $4 6- V 0 0 C UA a440
A. 00 43 "4 J.0
000X4k 40M 04 0640
I 43N 84 1400 .4 0 0 04C0 0"W$
H44 t 00 04X> 0004 o a (A
41 04U4 $,0P 0443 vf
04330 t, -" 400
0k x00 4 01 M ' 00 0 - do4k.414.0"0
>14614.4A. 01 0634360 430
14 [ 1
0
tM u
a 0WM:1 49 . AZo.
az
0
43c r le0r0N
a 04r: $ I~4j 4
k00 014.604 0a .4.
40' 0 01440043 0 a 4
44160034 09 -4G 0 .0-a
. H64>
ON 4 a.. N60 M4 HoaI 0
4J a00.04 I0 0 .460 O0141 N
NS c' 3 00 W~0 09
144
4404 A .m1 4 . J k4
C M4JX 1 rN43 r
ata " 4001 ,aac;
OW 000 0 t- 0$4 0 044
1440 0 6.14 0j0 14114N J I
0 64 CL4
N0 %0' MO>4 640 036495
a r lp s a 0
to
1~~~ ~ ~Ma
H
~
4
pa
01,~444-
U0 1.4 '4> 0
c 00.0 0i404
o
0
o 340k.O
4
c 'U 0ao
0 20.4 0 03,140 041 4 4 0,% 40
43 M20
N 40 .-
"4 1014092 4J6 4 3.
. i4 ".0 Is0- NO 4349 0. ff43
in ON 0 6 64 r.
94.)00464 0> -
. 'a .40 0 a a -4 4J06 tOM4 4
-4 k-'4*
6 - .c 4 4 " 4" Id 0
0 Z_ a c $ 00 a6-4 00.
$4 1 0 0 4 9. .00 012 0
c' j0
~
Z~~~~
4
4
A4 ~~ 48&3 .
0.0
40
4t qcn3 U- 0 04 .4
, 043
%4.4
0 A
B-31
4 0 1.0004 09 09'94* a0
0 0'.' 1
M W 4 ..9 A k0 xC 0I& *4 55094 %4 C2 0 ma0
0
0
-
A4J 0
0
0~~~~
. 0
C
0'O
-4: 0fW
$40
M -4M60643-
4c6cIMc4
00''
go ,4j0%c
we0
0-443
'a0 >
0
42 r
s.0
C4
54
a
-4 -443600
M9.4 6..0'49w4$4.0%
13 0 C043 1>0-444N4
"a I W400.
V40 w W 2 'aU1 6434 g-40d
43 U20444c 3 %n z 121.. 13 e -40' we- M.434-
-6 0 4 D.
5400r. H >, M0' 6 s 01WW
CW V)
jC*9C90 41'-'>6L.40 0 4544 A.
>. V2-4,-4> 1 C t, a k .
O 000 r. 3.6.4.0w V -AC >kMa00V2Q
005.-615 043a44 C6 m0 C.' A
l
11 IN 14 rE4'V0A0 4 G0-40. 0 0 0 ',M00' C 'a '- 344 50'- 05C 4 1.40-41 .
c 4. 4rM04
-, = 0 c .6 4 1 0 04 30 00 00C) 440 4
0 L 004 '400
0 6 043006~
->,oa. 0 CF
O0W4jV0W
430'6
54 4 431a ~
$00
5400"
-4 0
0
4
G44-C5
043
U4C9 w 5
! >34
00
4. '
0
W 04 004. 54 . ,0- ,a' to 0 0 & C -'.40 00
0 410 040 0 V0 ON 6-0 '01 0-4 k 5HU4- H - 4 P W 'W0
0 04 o 4ar4
C -
0
%4 0
9to4m.;40.1
-
%
(40C00
0 to00
4j54 W 00 0
443 50 0 000 43jr
~
S00.4 Q W.004
-4600 40 -41000
5 030
>*-4 4 -
440 14'
M-. 604m0
9 09430
m
0
ov.~ 04
44
0 o8
O
31
N
090'0 ,0 .4 -3014 4 W606.4 0 0-4 -~ a-
6054
'4-4No H.4-4.N 00 4 4 1 4 4 >-4$46 0 OC '49440 Z *-4 M 04 50 40 0
a'~. 044 . 0. A43MO N0300 0435t-46950J 4 540.0 0 '00
CO4j. 435 '4j H '0' z
19O 0 >C 4A 0 >.56
i 0n - 0 0i j
>.0WC 04) 000u % 0 .4 cHM "4 $444 A-t a 0 5400 W40
003 N.40a _4 M00M 00 Q0 00,000 k0Go
0.O'14 'N43 0''0 0 c. 0 m04. 43
0%400 43
k .C 04 0.
tC a4 .1.0 0 0 'A 0 &00 -40 440 C 0 -ik- 0 05 E 00 '-4
MO1O00U30- . 430
0 0043M
N-0 4.a 0540 4'.>. A m -00'
446ID 6004 4 J0 004 040-3.4 &_: Wi * 0V$ -t 00 -
0 41-2C
0 .
000W
06V>
in ix 0
.
.)w
'rM
0 .40
C
'.9M
0..40 4 0 t4
C2 144'0 M-1
000. t0'
so'
0 M
*-9'C 0-4j .4
0
050 - .- 0 0 Q 0 ,
00. 6
'ar440
50 r
0
C
MM
aX
Oa4.4-4
0 04
o0
004
U
113 .14 V0 m 0043 C
"4 454 C mO 40'0-4 0 0 00
'0 00
w& 0
00.
)1,
. V
0 401-0"
0$4 >0 4
:0 -H0 o4 .41Qn 0 0 r '
0W 0 009W4.4 X-4 0
C u
U)-0. cc 00.4 054 0 4404 Co 0 10 0-' 41400 0 a 4
n
H -44c 0 54 0 9CC;1! W k,6- 09. .4 H 0.0'-. 3 4 08-4 044 Ow
0440
0 0 ma aC 0 0, 0.00' 44) 0 0 0 OD~40 9.4 '044.4
04C-0 to4-9 01> 0434 6 0 404J 0'4C4 C 0 4. 0-14
00.40 6-
0.4 0N a4> 0 e.4 V
44C4 t 0-0(0 - . . AC.W 04 4.40t
040Z a> j*
QW
c j C40 , 04 0 - 6 0
r U)O "4 C
aaC
c54
N
014
M -4 -00
toc t
M0
0.'
t Na 0345' IM 0
N
00
'
0 0 03E 0 4
450 640
14~ .C4 0 a . - 4 '.00 44. 0410 44 0)
t & o~
0430
0 0 000 r
0 ,
O c 0*
7jC 0v 0m
w40.400> M40 00u 04
4C C0 0
0'C0
%
0 -A 0
0 10t
0-0
.0 to4)0 0 -0 430- 34 0 1 04 5400 430 -4 00 0 03304 00 t 4. 0 'c 00 t
vuwv VAV 0 430CD :05C V43 0 6-1 0 9C -98G000430 430M40 43 C > r&)0 004
N4N-4. V.4 U3 0000V4 4334340g 00o t 043 -
M4.4'00 0- 44 C1a 0 0.0
.4~~~~ 000 4W43'--
040 C 0-0 (A 6 .0.'-0 0 ,4 .45436
0000000 0042
019 6 -4---a54~
0.tr _ -. '0*000300. 08 -4>(40 C4 M0040C W0 0A
. 046 54C 11WI C 0 00'.- 0 W Q00C Q944 0 "1 04.' 3.
4) o CM 0054
U 6 0'
N5 0 W 0 '0 0 FA0' 0 0Ck0 E..0 6.40
C454.-
'01 MC0 0 0 '0E 34-
a'a$ o
4 . r0C 40 -4 M 0.4 WC-W4 4 C
In V60. C r 0ar04- 0 -m - % VWUC
'4 044 )C40M. >
0'I 00 34 0 0' 04 300C
544009' 446 6 400 146oNW 04 004>0V.C
44w,.0 00r.-44r ,W 0 0k,4(4.0 0 W V0 04' 4 W00 0
cC3
4j0 0 C03. V0 )Va0 0. t 3.440 . r.04 0'00C
CD94 0 00 c. 0 g.JM 54 " 0-
343 4000431 0 '00
N -30.0 '.9'4'450.v4 w0'0t 5443 04 044j0 0 00 C.
t 443Po00c a0.054 16 44 CO '0 06a 43 4 5454 0
10
m"4jM Ca04C3W 4 0 5443W NI - V
44.0 I ,40 0 00
4c ON44 . > w 000 54031 .00 000
44.43 0440, 0
paz0 1C00434 0x 40 0.0 0' 4403
-. 005 54 C.004$
0 C
00. 4'0 044V ' 04) 0 r 0r 5
M
40 94' 0 04454
4440 C $44440.C 0
43 435N
V4 r0043 f43 0. 43 C -.- " 0 X4 :. 004 446 J -'454 0
'- 04
0C4 0* & 40404 C 000-'64
V4C 0 43 aOW 0j
4J 0654at
064 .4'0 0. 0P0045C4 0 43.-I C H- C k- 004r04
A30~
064. 4 COCOON- 4 44~ A990 -0 0.4 00 0 - 0 0 C5V q; D
,'4 09" ~4 .~404' 064 0.0"C 5.462C0 r. .430
0'.0 . r0 1C.-
T 00 ;.,4 A4
0 -A V . 440.0a0
0 0.4 43 0-4 04.(
4330.
04
'0544
. w'Z x60-0 a
M0 0'0 0 a &
0".4-m OO 50 ;4400 x SC05 030 .O-14 -9.4O
.0'0 C00 1
or.4 4 13>434 90 43 a 00 5.0 0 00 v0 4 C 00 1 -
NA
a450400* .4 0 k94-44 w. 09r4'%. M 0 = kQ
000 A 93 140%. >44.4004
V44 mu0P C 3 m a4(9045
0 6 0 >.40-
0
ON
-0 00M . )0t
u40 '.0 340CL t'
2 4-, 1 n ) %
.4
054O
44
r6
00
a4"0
~0 0
00. CZ
9 446 44
-11
00
r-4%.4
O 0'4
4 540
*'04
04354 0k
0 0 M00-
5
- i-PCD 054ak 0 00k0 0 00 40 6 0:0 09 1i
V300 X-6 004C 4300. 0O '3'40 C c-44A40 43 540 60-QA 4 I
"446 0-44 O 0 0 43630 V 0. 5443
-00 $4 430a5-40 0 0 43 a 0 0 435k44 4 4-4>I*A3.640
0.4 r 406 5' 40 0 004 C JV 0 V
544
04144. 1 4
.0 0
M.4 4 A4V3 40,C 0 '7r 1 5 -44%4 4 4 04
In 0 0 0. 0 0 .0 4 4 .4
rz V~i- *3~4
~a ON40>
-1
4 4000 900
05 0-450000a'
6-OC
M
'50
OC
0.1
-410
'0CV
,a
05A4
M
54.40a
0a .0
.4'~
t I
it0 a4' r. 04
a.04 C9
4
621 0 r, 0
- 0
3 40
0
-4U 2 41
3''03ru 2 00 "40 50 0 4
I ~Ii~
~ ~~ EE.. 4 3OJ8O4
HOC....
44-4~ >14 0V... 0-4.)0
C >1.. 4
.1 .04-40, -110-0, 043 V )00
00 C.CC0 0
004 00
C '040'40-
04.0 00C
~ ~~~~~
>' 00C
0 k4-400.
0
W- V1
IM00
40 A)0 S..0%O- ;4- 0 'o c0
C
%0 aC 1 C4 0) &t~ -4 MHC W001 4,a0 Ok4J
e0
V
43~~~j
C400
*.-4C
C4
0
000
4.CO4044m344
M-~
-.
0C.0
140 0u4.4
4 w
0
~a 10
o
0
OC
~0
04 A4k 1 m-
*~~-
* 43
00.4
-C-.0 ~~~ 00
.4j>
01
>.000
001)
0 4,.4
C0.-
0 WC
1.4 C 00 004 r:C 40 U44
0> 30 u3-'
0 00 0 1A 0- 90. c* M0Q
04' %40 04 4
0. C0 - V 044J
44 em.4 -0. MX' 0'0 00
2
k0 CV Wa 4301 oot4 o k 0~ w C .50 w0C0
O 4. 0 40 C C13w $44 0 4J4 0 34OCO4004 1
0 4,. (4 A W0-J0
0k0 k04 0 C 4 w 0 0
0 .20 o C0 '0. U 4.-
) 0
1"4) 4 -MCO 4 4E 0 04) k
0. 1 .4
z .0 0- k ( 0040 0 0V0C-
~-OI
k0
I 00U
0
W0-40
.004 4-04
,~ C k
J 040 %-.4.r 4C
Cx .10
0
0.
O0E 02U1o4 0 - k 44-4.404 0 00.-A '0 0 w . 04)0 4.
00 VO -0.OC
k0. E4 to0 OF. 4) Ck A 0 --
04 fl.0
A.4 C $4 HJ C O C O O O C-4wk C
-4u0 10m w 04 0H 043 > 0 r-. 0.-4M > 004C 43-.O
44 m Cz0 00. 00 , 00 4W4 4 0 , k0000. 004U
En k w 0
r-r-.
~
C000j
o 0 &- )
0r
t .0
. 0,
0 '.
C
0004
0-44 u
w U)k-0.
g
0-0
o-'g
V 0
0
aO"(a>>.
'o0rk. I .%0 -4 z '0 Q0OC 4'V0.C C.4g-) 0 044
4- 0 2 i C 00r4 3 4 .2 4 1
C40 )g 4 Q $4 W-0 C 4J
OX
:J. C C
0.-r-
E :3'-4 -4 . 0. so 0En4V 0 %DU - aOkO)414
0 : 0 00
300.0 eI c r?0 40 0k C w 0.i n >.-4. '0
00.
W >P*.4 - 0k
4J J040 >.Cj( 4 jC. 03C
j k0 .C V 4 01
F 4.H 4 -C 0.
* WVCr4OCO~
4-M En4 r tq 41-00zV ,- 0 0 v a 0x Ck4
5 0
O 414 $.4--
4-) CI~0 04- 1J.-040
r4 C 0 C 4-0 C .. >.0..0 U W
040.44 -. 1OC00-'0
w 02-C.4- C0.4J . t C 0 w 01.C4 r3Ok0 k0.40V C
k43w4i-4043 044.J0: 0 Q V'0-4
k.44-- 00k3 0 1C40w .0
P4.0. E a-.40 OCO
0 o .4go.t43.I.
k k- 0d 4- m 0-C 0- C O 00U4 k3
w -44 0 1 .00q nC O 440>11C. 0 4 ) 4 0 4ul4340--
? 1 0)4) M CO-' -
00 0 C0. 00 V a1.
0 a>-4
C0 0 43
NO-W
w 0.
00,
) t* 0034
04 kk40040404
00) 1 W 0W0440 4C003.03
Or 044'4 4 00 $ V
C -
k40 0 V0 0ow43 0>0 3C W V)
4. co-C E- WU 0
0 0 C 0w-a C A WC0 aw 0 C w 100 OkxR00 r
4)' 4 3430P0. 0--4-
k-
043 Of 02 400 00 w o w E C**0 - H *04
U4 2IQ- -0004304- -. OkrikC<
C
0.0 -0. -4 &k0> W0r1. 09: 0 C .0 4.C Cn
04) ) "4 . 00 to 4) uQII 1.3
H .1- 0C 0 E-
43' 40040'4 M44)4
0 0 4
0O4-.4 04C4b
0 400
44j0 w4
00> $-,40 r 0) kOO v 0 0 9 .V C 04 C -o
4
CD~k~l
0)a4 -0 -. C 4. a ;>0" 4 040c to 0to go V
0~ 0 (a V C 0 rXk0 ,00 4200 S 5
Go C 0 04t I00.0 10k u3 0 U4Q) 430'
0>. 0-4- ,OD r 40 a >Oq 04-C a0. 00 COCa
0. w 0 .0uM- >, 0 0 0 > t
043-4h.-040S&
isar wO P0C.-O0U--4I4
-. oro m0 40 C Cs . t o0kC a
> V 4Vk -"4 CC0 1 0. 00 04
-'4 o0
ink l-'' U)- N1O >.H 10-0
40
*I)
.C
~ 0.C
0C soC
04
OC
a 3
toa
0
w " v- . 0 %M W= - 0' 0 0
it >VV00VN
V) C 4 : 001.4 0 C w
0.0A 4144 w 0dC
0
1
Cr--4
3 kU J0
-4)4)kEC10*, -4
oC C
1
: r j-
NqI'-kw- 4 N o4.D-4Q0'-41bi$ 41.
144
* O4j a4
410 d) V I4u w44.40- kA- o10 0E kCkJ
0 J Q= 1 a( 3. k - V
to0- >3
U.k kO" kIVUVV
j1
44 C thP" - 4w 10_0 0 C0 . 04
B-33
.4.~~t o00440000
kzV04.4
4)0 ~~~ >4 -'W ~ - a40.VaC$
-
C$4 c4 .004 0c4. N C
0 41 .044
aOco 0 9 .
0) > 00.- 0-k . o a .44)44
NsV 0 'cc W
0C 0 n 1
a O rlN a0 4) 04 2 1 *."490 11
0.-S a in r4 0 .~>0 00
.C H A44-
OCCwf r40
00 0-A 0
-4W- 03 . 4-.4> . 0 -4
<4) CME
0H-0-
~ to C 1
C c4 0
0o CM c -.04 4>O 74>
$4 HH-
&0 00 0 0 04)
, 4. 0 :CC-4
00 '10
*.04>
4)0
-00~4 4)
-,5 4 0001 00
0.0 0040 . 0440 O V.
-4 4) 0)
1%4 Cl tCo V44 "1 4NC
4j0 04 v)O 0 a c -4 A MO-404
0.$
41Ofl$
0 to
C4 0-4(a44 .-S0 *0 00 a 0 0 .0)a
N04 .) 'I) c4 0 NO CMVAO 0 6 -04 )
O'4)> 00 OCO 00). *j 00 00.
to~.4.4 0 W 0OC .e -a 0 0 9: 0E
N 140 .- >00 04J 0 M
0. -(14)'l00dA
0 $4 0 04 a0-0 41 0 44 0 0 0 0 -4 4
HC-4 W0av C* r r4a l10 i
) C)r CV 4CON 00-a 4 O P0
44) r0' V 4
-
CO $ V,<$ 040
0> $0 -0 r0 w.
. S W
4< *O'a 0.X OC--404J -SC
-40000 4 0 0 E--HC 4) 0 0
C-0. 00 0
.5 H C
*44.0 00Ma,$0 00g. -. 0 00 4 Q,-s
MOE-.-- Or- 4- C
4 - HOcM 0) 4 0 $4 N0-4 00
$4 >-H CD -.0U 4 ar
Ca 0 -40 . OCIV4-1
0400< 0)4) 04 sO HA V 4) 44. 00OMCOO
'-0 A' 0' W.-SO O.-,-SrZ8CO1 004) 44)04W
E4 -- >0$40. 0, 0l $44 CO-M O. 0
(is
VC4.- - 0> V 0r.CC4/a2 004 0 U-14 C0
<$4000.>. 04 4o4 0-$O 04 r. V4 a$OO
0s
0 c4- 4) * l QO o'D 4Z31 4) CI 4) 4)0 020 x C t4 )4
WO-0 0 0 0-4 4) N 44 -040 ,m 00 06 V-M M04 $4 -4
1- 0 00-4
4 VOWa 04.00 4A Z0 . '0 $ 0 4 44 V00 -4V
0C
4)0 >4-
04
0 >44<
V. $4.-
VC
04
0C
4
00 rH-_44
.0
r 0C m)0a
00O 1
~ 0. 0-
4)0-
0) C
05-403.F)H - C 0 0.,60
40 O 41 000 C 0V. 0 010 V4
00 0.0-0 $4 OCGo
$4 04V 0 = Cm 0 0-e4 k m c000-
9 0I-V)- 4.0 0400000.V N'; .- c O v Cv4S- 0 4 -40
4Z<E-0 0-s 0CO V< V1 -v- $4 4) d
0000 .
_ Q>.040 1420- CVM 004-4a .0 0
= 0
4 44, V,~40w4) w~ 00> 4j
<0 0C04V w00-4
OZ 0) rS4 00 4Z0$ -- o- 000 00S $
4,0 CH. 00 0. 41S W-SS0)0 W 0VS 0 0C0)
44 C44
T0.0 0 N0 0.-- 00 40a 0 to 110-
.401- 041 ,)0 0 0 a 3
.4400 t . 01% r.$$44 44
00 --4)0
z "-C 1 0I- to= - C'0.4 -to N0O--S0 001 0 414$4 M44 0M C4)
-
9 cO
~
0-4----44
l
.
V<)0
- .. V3)-50$
C
0
- >~ <W4)
zOOc<
O 4-
12
u >
0 $1 4
0
4 0
V 0
4-CSV0(
to w
00 z U,-- -S'C I U2>-4.44to>t <0 a000~.40 4C' s 1.W
W
c e . 4)4$
0'0 C.
VVC00V% 0 0 w 0 000> .0 . -C :
0 $ 00-S4,$0
4)11. 0 0) --C 0 00- 0 4 NO $4O4.50 4 )-00 4
0 04 .0 0 4)0CM-
0 $4 NO'k 0$444J 0 W -$4 040
IX,
$4 ID
-VC $4 40N 00)4 0C E4 04)4
4C ' 4. C 0$V0
0.0M $ 00m 00 $40. -4 -4 000
cc0M'-usC 0 C 0O a w4 -. a'0 $4 C-4 CV ,$40
w4 N nM 0 00.VCID
00O O 0>. 0 k0 $0w V Cg $4
4
C 4 00410 $4 C--S>.M0 JIC0 9 0 34 D .5 4 )
30 4JC0
,4-
0 4
4
04
OCD
r
r
Q,
E 4)-
a,, to0CO$.c-
$> -
$4 TW N C Z 011- M.
$400 a044c
1.4 00 1.2jH 4)4 0 4 $4E10 OZ -S
= -OS--0 00 a 0-C 0$ >-00 *CW1
4C 1 t 0 00%1W4
-- 40 0 ).4.
M 2 -0 04C1 z00c 1 0 0 j
41 0 4J00r4.00 ko. 4J44 l 110 0 1 100C 00
o'.'00
.4
044
0
W00-
0
4
04
tPr. 4
10%4.d.A. 140 $4-
40.401 =1 ,C
CO
W
in 4414-.00C
.0.. O -4a
4 $0 01-.00 1 . 01- M.-04 0 0' 4. 0
.0Cr*0
O4-.04 4 -40 8 00 $4a-.4
00, 0 C 010
G u44 C
014 00 4 00W14 14A O4-W C
U- - C-4 f 04
-,I 09 C:
q*- 10 40 - U ) 00 0 WO9 3 r.40 WK 0
O 4 0 0-4 J>. 0 $10to r. W -. 402 C.
a 014 0 * 0V-A4-M >O a v-.0 .% Q r- 4i4 0 09( 4 0
fu0 -AC9 0 'a01" WCa00 W=0 0C 0 -( 1 0
-0 c -U 0 9 0 $4ais0.410: -4 r. 1 W 4 4) 0-,t1- -H 0(1A 00
0, 9 V14W 0 0 01 N
ri~ noo 0 0 00
6 54 C n-o V
0'"'~ ~J 4. AJ~4
-W 0 t.0 0 41
10W01)4 0 N0
3 ) )4-
.4>.iA 40 -04'-C
0 0 4 40 V r
0-400 4 Q CC)GON'40M O0' '
Z 0--
4A'.40 kA00cc01-1404 I' z4 WCC 00.400 (4. C-
40
(/2 "4-44)40
*CO -1 .00
00'9VO
0 'C -too 01.-4 *0W" - 0 14 rz 14 M~ A--4o-m
104- (0 0 010v$ '.-04 0,- 9.0 00
Q4 SA..O04 040
M 3 $4 u (2 0
C.CU-40 0 -aZ.0 a0
P0C
4a $4 C 1% 14 -.4 1 0 0 44W4J 0
U,40 0 0 r. X40~
W V 0 FA IN .4J M40 0$4 4 '00 0 14 J > M- 4-'
r. c In - 0 1409,4 . 0 V . 00 14C
0 <&MOO c 'a4 c U-1 0
W Z0 N 0 0- > 4 9 0 00' oO0OOV0 0 Q 0 0 0 *0'-M 0
j0 04P 0 0 -A win. r. OA.
Pi WJr -. V-4v a 40a9. ac0 4 4 0: N440 440 00
q54l
0 0 0.CD4 X. 014.00 m 0 r-40 mw $b'4C0
a oe 4 . o0 j in 14.0- w n t
1% c -49 4 -M V)~a
OC44J 0 01- 14
000-'o VI 41214 31448> W170 0 -1 44 0)4W . V
- .. 4 '00-A -1) V 1 ). 0'Z441 M 0 Wo44
.. /C0 0 0V--4 1.0C
4 4 2 C140 ' W ,410 00M.. 0 jn -- r0 . r. ..0w0a 0
nWU
C4- 0)' 9 CO 0 0 0 A0 I )xHWCk M" 10 a00
.1 ( S.0 xO4 ") -4 W04U 0 004 -00
t
az 0 0c 1a4 .4 .0 OV0 $4 -'4 -49 "-4
W 0
Z15Cz 41 r-. 4J 0M r.MCOC % t arc
0 C 00 CCf, $ 0 -400 v oWI 4 41 a va c
u o - *14aa
ME.0 2 0 Q- a.0 W00 S00M0 Cz 4> a . $0
0 0.4-'C 4 .~ in1 I~ IC'-W0(40W 0 000 01-D4
ga 0. cmco 4
-0.. 0 0 in0A .9 W14 0000
04 IM0 1 .OC 0 0 14 wNo094i
0"
00
14140 $0 -4 0)M* 0 4V 0 04 A. CC 0 0W>.1 4
W "Wa
4)00W
*4J'
0
4
-4 01
44
0- WC-) 4 we m*o
C'O.
z
'4j~ .
W0-44))W
0
0o - 4
4
000. a0
00
004
0oOW1- A4'o-44
v >.0
0.14~
4)
~ 0~
>,-'4C 00 4)
A
Q
(
$4A 1
00W
Cf
- z $4
Cu - C-Wa0'I 0 0 0"0.-A00 '4 '4O'C
00 - CLA >.. 41 0W 04 Di 0 0
0 ' 0>.6- 004 0 " 01 w 100'W(4 0~r 04C10
00 4)..cc4cn
0
A0' 04
fu 0
0t.) r-'
CO'".0 OC
S4, CN
q00-
D
"
0"to4
V-0 Co.4101- V900 31 cm f. 150.
a
MO~9 0 1400a-A14
0 0 > 1 -OOO
CO0 j 04 -- 490 -41 W
-40
0014
V- -(D"n-%a4 09 r i- >I 04 a0
Q V 04U 0 M 0 0, 4011-0& toc
*-4
4)
~
E.0.-~ 0
1 0 41 0 0
~ $.
0 4-
i 306100
00
9
(4410
.00 000
0C
C0
-0 0 0I 0 4 01
C vl%
.
14(12 0
('. O0 .4 4 004
0 C
N o4r.
10 AC S-
0)..-
.0~~4C&&W
0 C-
Ua. Vv-Q
" .0 0,jS 0j 4 It
2 .0 04C
4
1-
1 U&
00
A 14-
C u-40
9 0'
P
'vO
> 0
4
IE'44v
, 0 0
I
10 V~ a -4V 4 0.0
,0
C. E- 40 1 0.-1A9-.48r. 0-4~
O" V 024
I ~!V00Z j~~oo
a r o 0a0
CL0114.41r41.9-
~
40 0 .
B-35
20
ro 00 V 10 m'
0 0 0 'n3 400r0 . 0
a..
-a . *-u04 I , 14
-' 40.2
"4(-1 10
'a4 ar
1:.I' 41 -4).:01:4 '0r 0 12.0 01Ma0 .11w0.
0 toH0)U H0 *-4001 4 100
W0
0' 41-
100 0 4 04 *4 0 r401
CA 4* 4)f
0c 00.
A 0-O'' 0. '(. 0 0 41: 0'-)04JW - 0110
0. 0 440,# 0) w0>0
0 10041 w tr w w 0 C. 4) r1
0M1 00 r.141 -0 c~ 0'0 0 1) 0 H O .
w 00 a 0 a0'0 m 01m14) 0. w
0. *-).
(D0 .q 0 0 O00 >.140 a- 0 411 W w1w41.-.
0
0 0 Q >&O >1:::1 Q
0'(V$41 00 1 4 0a.. 'D00M.-o-
0'- 0 Q* a E, 0.
M--140E: 30 to 10:--00.
1: l. 00 0l *-
m )0.-C -2 D.1 40
r=4 0- (a
0-4 m 0H .
k.C 0'h Um0 , W4 ), P. )
-m4 rO G 4jO 01 * ..M. r~1 4w . 40:a 4w
4 0 , C)' : 10 10 M: 00.4 10 00 1V:> 4
'(I) 0 "I . 0 - 4 :0 4 - 0
O 0 ~ 'N .10'4014
40 ON0.:0.9zI
0
0. -04
1 )4 00 04 00t 0. C-1 11
0): o
044 4
~ 0
to :1'o 0
00
1)4 00. 400112 4140 0 1:m U) r10.10 0'a00
.- 411 *.o0i. 0. fa 0012
V- 0.(A W 0- 0 -
41.-It
44 14 .10.1-14 00 Ai 001
E :40- w -c~' .-.o-o o
04 P 0 0 44 'a >:1~4
o,-4:a,
00 w 41 w- 0r1 u0'01
-H4
V1: 00N .0 V 0. .1u 041 D 14 0041 w v V1..- c V110:4--
w:
00 041
$.I r. 4)4 m.1to C) 44 1 0:.I.
r 0'.- w4) 000:10 24441 r) a) 0 ) g4 H0.
.-
M 0. 0 41 1::00. . 0 W
*.
0'' 01 (d . .0. a Q AUN
-4 0'1: 1 1 'n0)O
411N -1C-)
o -'
3.: 0 00o 4 M 0 000' 0 c w 01 L,00--Sim000.0 .1 . W.m
000 0.4.0'.0 "14C11 ca 0
1' -C) $0.0 *-IC 0 )V - 0.
03ax 14
cc 20040I 4 r. w 014
0 - 00 rW 0.4 c E0-4 *..
W.0
C4I)-I. 0 w4' E. C mX00>E. 4 UM
w4 0.41 1..- 01.0 4
iv ~ ~p~ ~ ~ ~ ~ ~14~ 00~ E00
~ 0 ~ ia~ -11
N410c4 n2 .'0:0
a).
-;-U
014..1.1:0
r041
04
. r
40
HN >.0 C( W0.C: 4) 1: r. 0~ > 04 0 0 4) Qc 41 0 000.0 4
11 010
C c 4j-' 140 41 411141-.u 1
V' 0 0 x0 1 141
*.- L U~0 W'
410 W-4 Z'00 - 0 1: 0 4W UN
0I10 0 W 0
D0 00 to 000 V2r
0441 1: 00 " 00: t
41000v
Q, 1 01 0. 0 41 0V 4144 4 ., U-1 00 r. r
0 0 0 4 R 0 V4140.0-- '40 ) a, 414
V V0 .- 41 V0 0r
<C10 W :1 W 4-0 4 V 03 0.0O 00 )4 d- to0 0 . 1'0QI '- - 4 )1 .
1:0 0) 41 411 4 414
04-41
0 V
~~C 444)W01
4- H'
M10:
1:4141:J
0E 0 0 WN>
.41"< $0 30 c 0-20a,"0U1
<Vc4)0 V..~
w::> U 0.212)C4
) "0 W
41 0>1r 0':414 40 : 0:00 ) )4 :3(c
. wv>
0 01: w .01$4 "1 0
r-11- -0 1::co
0'z'
0.4r .- IO(A1:-> 1
00to4-4
0 0)
41 (4
040
0
-I
0)00
004-
>
0 0
I"W
-9.
~0 u 4)1 P
H00
0 .01 0
$00
-- 0.101 >014101 w 1:4100 0-00 -0 0 4 '0 r- 04 0 V
9 (z- ' 041
I c40 to1 (a>V -.41:14 0 ' :o'I WO14 '0 - 0 4J W
'(:0 '04t
x1 - 40.4 w 1000 04-H 0) V' 4>..M4C
Z4 441
)140
0'40
* 0441WO 00 4 41 14
=- (a000--to -4
1
40 0114
*4~0 $ 0
0 014V ' 41 I)~
u :0. 0 C 004) 0 0. 00fa.
00 3 10 w 1:1
q 1)2
1:1 3-u41 0 40 M C'- Cl4$1V4.r- ) 01 :301
13r
4 CID
14 C *) 1: 01 C:9400M 1-40 0H 10
004t0(a to *1- 0.5 NH0
0'( .v '4M' U M 0 (C Q W 0 4 004 0 0 k4) Ln0>4
41'4 00 -0.-I C0 r' C1.:14a)1 0 0~
4)1X100V00
UC 0
0.- 041:0 0
(D '0
M--i 4 00
o o 41)1 0V14a) .1 00
O VCZ
0 4 c14%- 0 r _ >CH 1 u: ~0. m014 l 00 0c >41 maw ~0.
0)4 oo( Ito >0 Vt 1 O I V . a 0ix 0 0.
14-00 . *0 44OZ 3 ~0
0 04>0 m 000, a
0- :0 0014%-4 0~ 14 & V
M 0 0 CO '( "O
X X(P44 V 44.1 t
"0 0 0 10' V00 -0 33 w & 01
11w4
C 141 m4.
L0 - 12vCI 0000 v0 m40 0 E .1 4 0.10 4-~ w 0140
0 '.~' u 41 -4 4j0-: r..0 w1-wZ- 0 1ak41.-I 0 M~
-qO0 r 'loxI0 44 m0 tr1c a 000 - 00 m14.-C ( 0 : gx .0r4411C0
v. 4 13-
0.4 '0N , 1 02 $ 0)10 N4r..414p4-
W 000 -40004)
A-(
Z 3 4 I4:
10 '40 0 4'4'I0 ; 0' v.44141.4
a,0lU 0 14:1 C90 M a,.
410 40
r > 0 4,-1 0 0 VA040 0 on .
0. W4 1444000' 0 0'0
J 4) 012>0411-0 OD0 14U )- wr a toaa4 i0 4 C -
140 ~&~
00 a1)4J04 Q& 4 9 0 4.-4 0 M
B-36
rn
.4
.CT300'
41 C44C
M r HQ
f 44)W40IW
(C
U41C V 40--E-C--
*C 0 )C
rI-00.401nVU000
A3
-. 43
u
-
U
(44rA
0 0 ~0
r 0>01 W
0
A4) C 0
0
40
0U0 0
4343
A W00W
00:>0
w. . A H,40 .C E0 130 V rCC-
-(a
43 ('340
01.-i w-41 U-> C-i C01
43143V043W0.43
14 >. C, UO'3. 4 w000
>-'4 a a )0 o- 0 40 to 04J V -U W 4 C. 0
04 4) ' -> 0 - 0 3 W0 to 04 $ c3"0 430 ,U0 ..-. >1 4 a04)-
*43
:301r040
Wi C 40 43 'a0
OC11CCA0
o
*-Il
001)0.lw
w0 14 0i 0
4 .
JON
I r-C
S~,
04-43 9 0
01 V 4-4-
aJ0>C410000.A&JU
W0I
W C00 - 0
r
0 0U (003- -V , 01 0C010 C30 C Ca 04- .003= 001U- a Caa
x Q It04 0 z 041 -to 04 k- r4.W0
0C Oz 0 0 1)--10U0 00 0 00-.$
o- n 0 W OU--i010 004 0
a4W0 > 1 -. Z W C OV VC *l0 143r0 43~ 0
U y-0) 0m* 10 00 0-I &.H4 w2. 4314>0--0-- 4C .4
*-i 0. I 0 4 4040 0 0. 400In3V ow014 -- 400 ' 0 f:r>.
v(l*J
o g
4U2CC V 0
I-
0
Q)0443.
1M
04j 4E*
0) -i
.44
?4C -,40. 1 OC:S.
0 44-30-.10
4e0U0 0-I .. 40.0
0 -. ~4Xto
.4 -(M C
4i-H H
CU0 0 0 -4
W0
(a4--1-
4 0) 4C
0 0, 0
V.- cc N
WX04.01'->
W
M
14 -C r
-r
04V3430 0.
4043o
(
N-1
0-.
z
.1 00
N4->.
.-
0 04
'D-4303
43010 *4m044-40 p0
4 =00'
430$40
:(>a~30
4
a2
0) 4100(D-
> IA-A4 004
0. 0.4 w00040
3t3
34 C3
114
00
4
Q.
go
43
:1
0-
0>.-0.
Q
-.- I
W..
C 1
0404
0H
C(44
0IM 'rUl 0404 '0.QU jC
01V M
04
0
U-40 a
43-44
0
0 V.3a
30
$. W.
04) 3)0.U 9U
.0 0
0 ow0 0
tp =u4
000
0
0
000
0 C CL
> 0 >.WGo0 0 tr4)14 4J U0 1430 0l4 V3~ 1.40010 ar0
M01 0-0 10004
,4
H .- i)O" M0
04r 0
-4)
Ua
01 03t-4 a I40
0(D'
0 ~ 34
00>.3 . 4)01--l
41 M 4
0
,0-
0-40>0
404.
C0-- 0 a
T)
'434w0rx
0.0)
w 0 0-
EC y 01 Q) 4)340 040 0 .0 0 r 4 a 0-.40 0 0vN 0 UC C- O 0 4 -i03
4)~
414 000 ( , , 34J- 0 04 C~3 4J 4j: ) C 0 4j M U0 0$
zo:0
w < zE-4
0EI rzv(y, 0(a
0CDC
VC 0 a M
>0.M-1 4 0 do0-0
a 0 k00
0'
0
44
OOZ
rz
4)3.
410 ~ 4) 0
C
A
43004.) .(a
0 0 404
00- 04)-
W NJP :3 4 4
= n TO-
W00. 4) 0) 4340 W 1 .1C 0 W -01r * - ->.M4-i- 04402)--l0> >
4) Z W C0 4 ' C a- 4 >. .,>'0 , 4 4>.C
W Cp 0 40E 0- 1c=
00 0 04 mU4 - V;4
04 C w V0 0-4 0, 4 0 4 04E0 *--4 a544j-0, ww
403--4 CO 310
4). 043C >,W4 0 4j U
030w -. C-I 01 0 101 043 0 0'04k0430WOx w34 -00 IOU)010v.. 010w 0 () 00>1U-m-.
0, 00 -043u 0 aOI 434r0)0 -i0 4C 010 =
CO -0 - Z:3:C S40 0-i.C C - fl 043
0.0 H-4 m IA a, w0 3 0 >10 4-C0 -4 0-- a1 to*010 0 tr'C40M040W4V4V.4J 4 044
M 04 U Cs3.)
- ) ) Q -4 b'0.-14 0 W 0--4 W. $00030
4--U 0 00 t 4.- > C-
a) =O 0 -.40- C0C3~
0 0A43 000-.I . 43-40$0 0
af~> 0 m01.44 0a03 43
0 >~ U 41 0.0 4 WU1' -4 0 30-i 0-404V-z 4-4.0 0 40 C3 A 0'
N4343 4
.01 C 00. C E '0.- 4, R.04MI (030 400-- 0 430 ->4 4 0 0 MC40
cZCr.M U) C0 -).- 001--4. -C4 3 v CCw 0 03 V U03VU>444 U304 E-(A
co4C.0-4~ U)AH 4CDI 43 C4 *.0 0a 40 o'04). 1 4 uCDo 0 04 403CU4 E0.-.-r CC 04 0 0t
co0 . - ' lC 4U-C01U 04) U310. ) 40 .0 4- Ar C4.- V - r CM -s04 - C C
NU4 4 n4) 0. 7 C0 M30000 0 0 0 0 0 -
M. 0 1 4j -- I0m3.w13-.3 43430Q3 C-i 0343
Z 000--r-0 0 Q 431 0 *H>C 1 0 1.< C C 4C0 -I.4X03-0 -)I T.U)0v4c 04) 3
cy 4 ' *w 04) 0w -00 (1) m 0 (4 a,
O 40 0 0 ' 4,r 14QO0- M5--CC00S0 0.Ow
co 0 0 'a. 030.I 410
OC .L
0 wC 0 00 04 w
U.0400'-
1 4) 003 0 >- 0 .4
.00n20 4 0 : 3 4) 1 a4 a( 1 aO010o , S 43 '0' 4.1.1, 0)
0 v > Q~ 0 4>1 r U43> Ai 3t C. > )C 0243
(D- 0 0w4."- -103
j )0) -HC C 0 , C4
0 0 ) 0 43 N - CO4 4 3-(04-I-I-I 0 O-
04-C -.4 0 4-.C U 4 --4 0-. 4 04
L4
4 0 0: V) w. V''--4 N(3 001
41.-
CO U 43 - 43 4 m04 r0V0.0 0t
40'0304U
C.(EWA z 0).0 w C -IC 040 *. 4 304 to0a) C0 30v -U C03,-4W u -
.401 0) IV 0--14w
.OZ-, 0 C
0 CU C$. 0 000 W
O> 43 VMC -10 W0 W W
00C 4 .C0 0.0- 4) n 0
-'- C n9 - 03-040CL
0-->00r 05Z 0 -0340 040HU0 4 4 -I 3 430
A 04 4 MZ-04000 0020'A C 0 sOW *0C)~4 40 C 01 4 CD
4 W W M. CC4..002 -H r- $10C SOC 443.0 MVr40 (.44 L
43043 0 0-.i41C0 W ;CUa-. t0C 0
0U -
U( Pi43- C2 40. 0 )020 310)4H 3 0 N <-U 0 M- 04>,C (U M0 0040 4.)
ta Q
-- 443- -I 43-o-l 01 4 iU 0 U40O(U 4 - 0040V - 'A4C--0'40 CD 4-. 0'j
1.0WD4 0 r w (aW 43 4 C m04r00-vI.C w M 4J0 M-J M U3 Q434d-W443CC--100'.10C0C -
0) N(n :--C 0 (D 44 - UC W .40 0 w C 1-Vl(4-30 - M'-X- rN-4- 0 , 0C W-r 0 V
40-I 43-M-V)-_W 43
u QDI 0 06i 3.-4j 2 - 'a V) M Vi40M10U--
M 41(4.
M 4U.1 U E ra r3- 1 0: 0 : r- 00004040
v U003. OCO 0 ni 4 0004)-
-- 0 0 3.0v0 04004.<.4-402 03$4U4000A00 - tr 4.) 94)Ur430a
000--H 3>0 1-I.)C 4.) to -A0'1W 43 0<D0 - 3--I-i 20 40 00.00a0.040s 30403w4a0M>w 2
z HV 4-.0003000W0~~ -40 0 2-1 l< w 1 0 0.433.IC0 $4 $ O-13 o-:3.=0ED
s-sOU.g)I
- 0 0U0
o 0 0 t-iU4043 0-04o A 0 >0.U0.4000-I4W 0 4J0C---V'
E-.4V CW40-3.0-- 43
l 01mix0-IC, 0 0-4 C : 140 .0 01 0
C.>0 -M 0 a 0 c 40~
0 43-C 0 V3 V. a30U4 V 00 0 43
F (0 -4400 C 9: 0 10C C- 40l2
W 0-4H3V 3..-
r. 0 . 0,u 0, - v4304 4 04 0- 4 A3M.iC 0
400I-I 040 C Q 403
9.64- C - 0
V)~qb 04 9 w J go V V r0.41 0M42 40-s
0 0
ZA4-if P3 - a40 43
030001A. 41. 0- 0 40 0.8.4 30
~ 0
9%444
-4 D . 0 V U40~0 43-.. 1 00 4 N 4J 4304O
N 03-
V) C 4 0 -- iV V 00W Vi 4 0 N 00W
0. H q 0340g. 0 a 0.4...
20 M0 3
Hce43 04
§ 004 3
:.40 041004Qm 11"
U
t- B-37
r* 02
IDr J 0) -M0 rW MM 144344-4.4M
U.N
4- 03.> 0 1 0 W
4J440 02 0 0
>1 r-.4. &4 23 00044'>
c 0 4 W 4 00 4
.0 "---0w 4-1 R~ t- 00440 to..
04-4Id 4)0 0-4 -a4-)4
o 44 t, o04 4 00 ) 4 4 - ' 02440
40 44 0 4M4
t 4-402-0-..0 3p 44000 to4 0404 o
U)
44~002-
440.
~
02
.4
42
4.4
43.
v02 m0r 0
0
in
r2~
a. >.10-
00444-
-A4m444444 r
00.
404.44.
r4-
14.
u
4- 0 440o
'i
co444044> (a--04
4 M
4-m 444)
2Z -. 4.4 2 ->00
0 0 41 0 .4.14-
a.4
4)02 404
4 4- 0
a)
4.'
2- >. -
C440
.40
4)-4.<
0.0 MI
1:6
020'a4
$4.I0,
S4 444444
u
0 0- m42. 0 A4
"- a.
3 2 4440 4 V
4400 0 0
0 a) L.A 4 .=. n H-1I 4)> 1
V4 2- - 410
4 0 w2-.4 4 0 400 044
4
4-~ 023LV .. r 9 w 0 o w
r- () w m4
4444s 44 04 44.4
2 4j
ix42.a . -40 2 J
2.00 0- 4 eq o,400244
'-04> 0
W 4 0
3>n4 4Q 'a<
442 a44co0 o<102
cooo~.. t 024-4 O
:3 VX F *-00 .4'4O V-40
w0 a44 )00 z U, 040 w
oo. W, .-. M 2 - 0
V4 02m 0 . co
14-214- 0L $44-0.4 Z $4 W, 4...0-444 0 0 ( 02 W a 0 3
'a 44
a- - .C3Go 4 c- W4 144 4 -
0=.4W - 2C00 <mr- 4>-
0 0200 0 2 4 - 0 a)~4 0 - 4 . 4) - - -0 .
< . 04 * .4-
. -4 0 .<-0 4 e 0. 0 2 > 4 4j - - 0 4 0) 0.
E-20 2 02 E . > 4 02 -
r
-
0m -0 <<
-
~
,r 0 Uf
0 2 W
0<2
0 X0
~
' Z
~
41)
040-
4-----..-00(.E
- 0
-00
2 0.n 00
. 4.0r40 4-I
j 4M20-W0a00
00 4 < 2 20 - 4 .X0
.0w
r
4-2
E2
>.0 U .24.
4 )
0 W.
4444
V 0--4
4 0 4 0
)0404
$.4 4: E ' 40
w -1 4o 0.
0 a2 44 1 W W0 .0
VI
2M3 444M 40 44 4
--- 0 2 0 W O.- 4- M4
>,04 3 4 -WN 0>0 44 . 0 .o1 24 4)
.0 '-4 X
a 0. 200
~- 4) 42 4
4 0
440
4)
0M
-(>
0Z
- . C.0 )
C444
44-
0-44 4
444
r4 10 4
044
4 . 0 . 0 0 -0
4- u-2 24 a-44 4.4. 02.4 to 0 4) u >.4
0 0 4 - 04 4 4 4 04 0 4 -MQ4 N 4444440>. 4 444 4
4' 00
4) Z 2 ., a-4 0.0 u0 m- 4
-44 444>- 44 .C 0.
4-0 344(0- 4 2 W44 0 -444 20 0V4 .441 " E4 ) "-
02C22. 02)2V 4 r> .0 0. . 4 40 2 0n c 4440 a) 4 4
0.2 0244 u CD 0.N 444 M404 r4.>
(1.2 0 0.)
4404 44- 344 010 .4 . 4) to 4)4 34
44
40 3 .-C)4.'4 4 '- 4 w40- 04.-4443U0
<44 4 0 0 -0g 0 --0 003
4 04 444 -) r 4
0s 4 004 34 4 4 ,- g . 4 4 4 4 44 0 0
00I44 4.-4-
s04404-4 . V40W.04 0 to Q
..1 :40)) m4)-4 U) 0 .002 0 44-'0 44 4 .
044I. 02 4443D44 ';4- 1 . V)34
-4 --.- 444-.
3<-) 0 .4. I. 40
4 0 w<. .4 44r M 4v4 4 1440 4
OD04024I.
I J-
u
u0,U0-4
140.4 00 0
-M 0 04-444 44
fn44 0 V-01 4-04-04.-
w
Q P >0(0 a a.u 0 .0v 04 w442 -. 0 CO Q .,a 0
V
MC* W
-1 -0
4E-4U'4- . .4
0 . r-44
46C.44
0 -w ?
_4C
~ £40444
C 0. 4j V a U 0
14 tm0
0 0
00 0 0 0 04 06 0 . O4 :0 000 Li 0O
co .'O'0kA40.0
-. 4 0> -144-
44 0-460 -VaCO a o 6-no0 a 0 aV C .
44600C£4 -4-4-C 4 ' -1 0 0064A0 C0 £ 60 -a 1 - r0
41N0
r4
0-4 M ~-4 C6 0-4> 6
k 4
0140-
j1
- 0
-,I V,40.
-. %.4
0 0 a -4
C£r4k0
44 > 0 r.-C a 4.-0
;6 4
4 V. -4 1 1.13-44-OD
P09 >0M6 ,0'.- .1 o
S4co)w 0 0 ' -U 00 -0i *4
C> $4-4$l0 -4- 10 4
x4
*.. 0 XL 64 r0424
4) C4-4-0ma 440 "1 0.4
306 0 k
OON'010C x 1
44 lr 0 0_ 66 Q,--4
46to44 046W 4 1 0 0660. M1
-
0)C 0.4 U
044 0) n1 a0 L X
OD40-s 00Hrr 6 601-UVgEn>00 rCr I
.6 . 43 '06 0066 06 .6t4'
w 0 0--M
E- . c CO't
41 4.
00 L, 3OIX
£064 39w -,0044u0Z o
-i
c. 1.-400 V -)'0-4J- 0. 04 V4z-V.'
'44
CD 0O-00
W4X4444C k-1 W C 1t N>-
E 04 44=
R £4 0, $4 0 '4'0 1-1 r40'4w 4) M0 4
£4 0 04w41'C 0 0C CO01 640. 4 -4I
w
00v 0'0 6) 0 R 0 0 -. C.0
u.
- w0 w430' 4 r.Q. .
0 1 U4
00W440 600
: -4 0U
6 0 0 HN
Co 0.U C3 I644
44r -4'LaO
0 10w -D _A6 r0.
4'66M4-
44E 6 Z40'
-)6 -l 0 36
t)4
£440 a 00i.-'
0 44,02 4 0 044. k
£44400
441 C 6) h4 .0 = 4 4 OCOEl -4'- '0.10
.4 . a606>o 304404a4X0)44 $ 6-. co 0-ame4C
4 u= 5(. -1 $. WQ -w 460 0 -2 l 0I -4 0 - CO 0
O1 C -l0 £4 ) * L 0 IV -4 0 o
CM -4
0 440664
0 to4) 410 >w v a0 40 0 -1o06 0- CO
O0 (.C4 V( 44.60 00 a >. CC-1o'
5 4)
04 06 0 0 04-44406w -40 004j- -4 V 641 44C
$* -4Z 0 w 40 444V 0 V 4.; -f4. -A
).l a 0 01
COC6'4C A4 0 0 0-0.0 a, mi OCO k M 0 60444460
0 0 044.) ON0U6 m. 0 44 00 T r4 g.6-1.0 -44430-
044 £4 C--r'0 4 . w (a5C w 0 ON IC 0 44 A-.
4-4- C6 3 0-1%0-4HC
QO 440 4 C 4 V4 44
C1 -.4r
.0 m400' 0 $., 64.0
464 .
(6442 044 o63 lid £4. 4C(n 44 0W '0 W44k.
41ra.a
40 C 00 04 ,a r -1 o m A -0 a0
60.9z C-
4400 44C 46C4 0
.0 062 44
J og -4
W. U Z4 f44. 44 66 00
0
Mw' 04 ,4-1 ,4.1W
04M 4 r. H0V 1 4'0 C
00V0 vU
-0>4w.. 00-0~04 44 0(00 4
04mf> 0
10 44 0- 0 1 04-44- s.4 90CmW 6 6 a 40a 0
060. CO 45 0. 04 44 4400 a',0 g .6
W I 0< V64
6) 0 V r a 440 C 641
0 w A 0'a 0
0 1 -418*.
2or .J 1,0 1 ( N.04 bw 41 40
V44
-0 ". E ...
06,H4.4J0 '44 '04)N
N0 - 10
60 0 r
600 01om > 04w ti. C4)a) 1..
10 0404
U
k,
~
a 00
%4.0
4
10
c
0J
0
oC
0m
o
r.0>
. g-o
0'V
V 0.0 004 c r 60-.I 0 A**- W -0A.0 VW
V VV 0 1.4 IS 06 W
0.4) V) . N 4to.V 04)
C3 -00 04 40~~~~ 0W -E 0O0 &4C C VJ60I 644 04~ .0 -100
0
-(CA3
0
>.0 0C a,
* 1 QOCI O4 0.0V-4 0 4) &0400--4 0 mo 0~41 ..
S..E0.4 0'4040a . t4) VC0 0.140.4 64.r_
00 C.in 0
41). 00 04 0. CU0 00 -'in 4, m -
-- 0 0 Nto 40.0m4- 4 .0 toa U -AOE- 0 n $ 4)
It i G lg 000 0 80 -- 04 0CC w 0 ru 40 0 40 A .U
r00 -4J 0.W M60 t)'0I 41. 4 0 $ -k sM40o -0 C L4A J 4
4)0rco 0 0 11 34 C1-6 10tli. 6-44- w C0 4) go 0 V w 0
a n' A- 0 64EU > 0) 444.40
C.. 4j
.4)j4 0)'
0o 4)j ' WC-- 4 W 0040(
0041
.4 0 4 C60.'0- V 0-- "41)>.4.
CA 00'4-
*44 -. QO 4 -44.. 0 44 4
6 A . 4) 0 0 0 >.6
m 10 604wa) t m. u W4-- r4. 0 r
-
00 I 60' H u 04)000 Ac Co4 V
00 14 --. 64--> C-- M.0
C 4CCO0. 4-0 w 40~ r 00) -4
4 ) C).6 4 -)4 30)-4.OC- 4'
O E4)404~~ A 0 I 1.460 ' &0El.40 00
Q- vC-4 001.a 0'140 14-40
0 :3 V) A- 00 .0$4494
:S u0' (440m )4
.9 0J.))FA0 * W V.
w.
toH0 4) (n 0
*C44
04 % 4
4) W66
4J
CO
V
C4
0
-4)
0 . m' 441
-0
1$ 0 0 >*.4
4-%0 .r
4J4 to
t M0 o
00
0
-
10
n o; I
46
4 4)0C
140
?A6)
0
m0
0.4
H0
C
E
63, 0k-4
6Q E 24 u
C4 C 94'4
0
W
.4
01
a0 0
V4.-14
V66001
03.4 c
c
U
014m
wU r. ) 4 '00 M-U0 V UO-A 0CIwe%0 30 0 014) C6w663
.0 V0
01404 )wI V0Q 0 w 0 94 H-4.i4I
4U
Or--0
1. $)4.VC 144
a46 >-4 04 0 .4 -4I r.0 W0 0~ V 4 4C) 00 0 40 &-4-4 4
6)4 M.01 4 0)A4 4 444V0 a ,A4 0 0.' 0006)4c I.C r4I0to).c
40
614
.04 1w
4) a4)6
V 0.V 0
I =601. 0 0.04 *f0
U44o-.
-AOUC
IV4 80
)4.-4 .40 a
V
0
)04
C ]0 4 -4)
6)X%4 0
*-0CN
r_ 4 0 00 14 4 4) OW A*- W 0 0 4M' N :) 04- ('40
E6 1.044. 4 0'-.'W
CD4 r--4 OD.4 Z() 0 $ 4%4 -4 ON-' 6>1 04Vt-0 $4CA Z4'a,4l : 0 6 4 > .a
( l 4 -I 00 o04 60 00 l< U 4 10
00 0 60 0U 01 0 $ 4 0V M 40 $A 0_ 1 4 01
wQ = 4 1 .4 A.,U Wo . ~.
0. 1 0 0 0 -0U46. 0M 0.0*D*_o 0 4) 00c00
W:Su~~-
0 AW( a -0)00 m.C 044) C
$4E.
H W PU0O.E-0 0)W44ed0 0 4J O0 W0.-4 W )4w '0 -04 t" E.0 0 .. 40
0 u0 *C! 20a $44 1)W4 4061s.m4)E- 04JCH
C Z'U 10 010 3.4
0 *.4- c D. 4 0 ow 0 to .Ocl' 40.4k0 6 6 4
)4.00N to w4) 0 a 0 0 0 C -0110 0 t,041 1dC 0
40, .. 4 - 0-A U 0 414 k) 0)$4 -4 :3 0 n ) a m..4 060 E- >.N
z4) c6 w6.44)1 4 0.0
MA(a24 * ) :1 -4)$4 03 0 .-.
4
.< 01% U4-.46. /M0 4) 106&)40 AJ-4A4 cn *64) -.- 460 .4)4 Q)0
c z40 0000.0 -. 4 ')-14 100044 1 00m 0 r4- E.00 0 .c 60 w z-cO
0 Hf*. -4 0.C0)4 0 020 )Q>4
a, V0 60-I
0. (D 0 0 44to()0 0 A-00 J - 4 - 14 4 J. 1 -4a,6 04
40tako0- a00 2 g 0 V)40(3-10E4 ?00 N440
- z 00 0
22
A. j4n)04U 04 600 0 . *0 4q1VW0 - >.0 60)
.46.04
6 0.-i ;* 0 0 C00 641191 00 vt -040 0 -1) 60.
u 0.
V0 04 . - 4 0 ,$008 .O I *0a
c0 it 4 DE 4 -C2 10C.W -0.0
4mw M-A14 40 - W o -.
t 0 4 C. to O- Co .6140)>,( -Ar C 04 0 -44
640 :3 0-.4 M40$4)0V M 6.4 1)4 C4 -4 CO. .0 4
4 Z 4) 4J- 0-4' 6)- .4 4) 0 2,4 0 0 4)2-. 6 C - 4 -0 .06)4W> 440
M404. )r( 4
CC4)' .0 066 000W42 01A04t 04.c w4)0) 1 4.40
0E-4-C--9HCC0 E4 0o0 0 00.00 .0 0 04 a 0 CD 0 )H c cc 0 x 44 > AOOm'64
w661w 0 0-.4-.0 4-4 W4 0U 01Ca > 0 w61.0 -A4001w4)0low
I- c f C4 Z M41w 0. 400 4 'aCN w E.0 of n
u)-Z 10 0a, > 0
C ) A4)Z
- 6)4.W4)-IJ
W w4)Z - W -4 W4J r.
6 0 4 w 4)0 4)60 0 4J a d
0.4U 0.
41 w FA~E w) cNE -4w6Q.
60 ~~E
U146 C
CV
46 N Cli 040)0
04) 4 .C U0 4 ) A -. 4J M4W4)0
w0 '0) H 40- 0 4J) - - 0 0)4
41 >1 C 6'> 0 . C 0>
4 w0 0.c &
C 0' 00W -4 0. ,0 4-40& -00M -4 4140
6 0 V'La-V0 4)44) CV 'La rj44.40"4 4V
0~~ 6)0 0 .4640 OC00 r.4
S 0 00 4 04)000 u0 60 0 %Qw0)4 - 0
w4 c c m -W 4 0~ 6nc3 tom 10.4 >$4 0
V4) 60
W C6Ln 641 -40 v 440C .L .014 '0
0 U -. 1-0 04 Al 6o10
a-4.-04 0.0 C
C 6) O
6.4- 41 cc24) '4 60U OD
-4 4 - CO
E- W 0 410. 0 1(A0 01 0 C0 V406 V W .>0
00S6O0 44-.0.)4 C -0C60 00 = V41
6 0 64) D0 0 ?A aw-I .46O0C 0 - V)
> c 0 > _ V)6 14 0> --H -A4 '> 0 I
0 go 0 -40 c a.04V0g 0 w. Nw0 464-M = I
31 -4x A-.0 0 'S)C -4~--. X 4 0o. OH
k4
0 C
u
s , g 6
r'
r-60 '0 4)
-41 004
-0 -4
V)
C
V00
C C ;-O0
~ 44~0
0 -. 6- 6
-4 w4 0'M 1 0$4)64)$4- Q 041
04 0.40 C)14040H
H 0 -4C 4 * .4 )4 a 04) U 4 4 ,-4C c 4 -4 4)4. 0. rC C-M
a a V --.- 44 0 .4)0> 0 0 tP44..4 4~ U)M-400 .40
4) 0 W00 (40t - 41 -4 10a 06 CO ( U V04 04A c--4
0> 14 4' 06 A c00o0
A 6a4 4 4 '4
0-.-.-6 O 0j 04, vv460
6 w0 0 040 0 4)
u4 604 C-) 6--4 404H.4 .1 HO w.0
&4- Em4GOD, . 0 )-' 144) &
)40 .4)1 I .06r-Aa4
60 's 0)%]C 0
0 .4- 14 k0 0
14 141
B-40
>. 041
0( m.~*~~1414
004. 00
-0 IV
0V>
.04 00
>40
4C- 00
.4.
4) VNwW4 34. Na
0 0300 0 0 0. V (a4 &0
4A0
c %40 2 00C -. V
0 044D 004
0 0 41
c 0-- r_4
424
-44. W T.0 Iff
.0 4 0) W.-4 C0 441).
n ' 0 6 41-V -00
:3 4 '-
41 0 a,
0n 4) mm4>4 90'00
00) W U:1. C
-.
r,-- c4
043 01
00'ja.
C.0 W
-W0 ~ Y m -
0 0
i 0-5fl
00.>
aO O-> 04>M$9
*. NWco >1 C04(100V-H03
NW W>041 u . (c" -
* N 0 rW0 .- 4o 4'
0 44OO. 00
01 1 0 00004.0v10.0 $4( E1gg
M$ u-4 2.. U404
'4 c Q.0)1. 0) ta 0aE-4 WV V 0
W0 (o $ 4 0 4.140 V04l0. 0
CV:3Of.0 0 a0V..c0 40
*41. 4 v 44OW4
0 41 _ wUn1J 00nZ 4o- Coio4.4X-I
C0m CL 4
41.-4-0 000 j0 14
50 U; 4w .. o 'V...
.. 0 4 .k% 00 tP
.400
W n 4 a$4 (a 4jC oo 40
;-4
M44. rm. C low0 VO-400c0W
W
0-o A10.40 . V.~0 094 0 0 .00
to A,4JC 0 0
0 v x.
-- 000 0-.>tV.> 0 .0 4 C:
H1 04 04. 0 0 (0 . 0 >10 =aU0 0
-4 'a >. 4 - V0-4410 W 0 0
0-0.4J IV0)4-
I0 a)O 41 04 U0 . 'A
41 W~
0 -0 0000 0001
-., rzO10(1 M-V3 -.. 040 41440
C 0 o)4
0
04* Orv, - 0 W--HgO0 0,4
to*- 0 . 4)
I O4.-W0
o< V 0 0 J0-C0 0 4J W 0
4D4 M 0tomV
4:>.U 0 >
" ur.04
c It% a,00 W4:-W
0 0 0.4. :5 4
t t'.041 4) -440 0 000.
0c - .40 W0-4 V0 0 4
40.4U)0 r-
N 1 V0eAiW
4- 40 4 L
4C V m >0 04a4 010 V 0C) 0
*a_ 0.t
C 0 00.00'2
.o Z3 V 00P ?AI4 ) 4 0--(dco
03- to - . 41 0.41 0. '-0 0 41 w 00
*-I 0-0 3 o-
wi4 a0 rr 00 N
to4. 01 (2 00 0 3C. , 0~.-.-
9
.4: 01 mO-m -4> .-4 -- 0 4
00t
04
. 0Q o tn e0.
404 00to M U4
00. 041 to 0 . o0o4.0 g
0 0> GO0 $4
0c331~.V C:1 40
-0 ED 4- C co 0, 0 04:0 0* (a0.-O'
4 00130 0 44
0 W0 -4Q V M
0
003>
1 r.-4a,-M.1.4
W1
-~43003o
0'1C 00 C4J
0 4
0ix .m0 .
-0V1
(10
3
-4 31
C,<4 0 W
O ~ 1 :3
o
4
nV
,0
u
04
M40 ..
N
r c
>
4 ~
m z to03
=1.44
4 >4.04 '0
0 V-4.J'0- 1 4-4
r .-0----.-4
E C W U0040 N.:
. ~-4
0
(1
to
0
04 V
0- E fa0 0
4
-4
0.410
) 0 r o m00 m-4000
3) 0-. 4-0 4 i 0 04) 043WM.M4 to0.0I -- (a00 0) LV4 41
M
- N. f-m . VV.4 a-HZ 41
0? '4J W 00L.0c
c3
L"J m W 0 0000W0)W 0
040 in to 4 -)0Z MH 4v00-o
4 . 0> 41>00..J. * c $>.
t-'. W 40.1.4%0M0104
O403 0 Oulu
. 01 .1.-.
'Da m 0. 41-40V X43 a 4M
E fa 03-'.-4
tWo
= .)toM 0 a
10'M 04100M 0 -.
44. .0 Z ff 0 CC 430 41 44
0 0 4-. , U V
Wo( w coU 0 >
Z m.,
.
03W
430%4 0 Z11,QC
00
4Wa ~ -H 4 .034J
3t M440
. V100)4V0
400r. 0 0
--40.4
L , 0-4 0 404 0.
040>.01v 0 -Z ' 040
ta 0 0t-404 0
-0 W.-Q
000r. q.. H
~
V4 > o n.. 0 4.4)2W.. 0 .. 4 a414 0 ~4
04 0- 0 (n
m a a n-W. 00. Q c1 404 >. - W 0N0.M-MO.L4 0 0 -. 4.
00: W W0 40.0 m400.-
m 041 o >.3040 -4so
-C0W Q V 4) c Lo5
0m -4 t 'p3 0-'-- 41 W4
WC 00 $ 0 .- 0 WI=.44 0.4.4
00- 3tW00 .. 00C .4 0 VO 000 M> 0'4 0) W04
0(a .04OW10 eq
10) >1 0. 0-
--0.- > 00 0 00 4J 4 r 0)
.0e 0 0 04-). 4 Ai0 zO 0-
00u c 0 %r C40
C.
44.- 04
fu Wmo (a .0-00 U4.. 0.0
fu 4J 040
41 (1)4 04.4
W .44 0 0.0
M
L, r''4~ c 0 C . a ) W.. 00z
- 0
00.-- . u0 003.1. r.. 0 )0'oj r- 4 _4
0 A = 4 0W 4 VW. 0 4 0 --0 000 W.~0
m-040.. -0V 0, 41~0 VN'
fa '
000 0.m 03 E.0 W- 0.4L)M. 0. 0Wt 0' k
fA u40 '-- -4--40W000h<41 0-1
001-4 4 t04 0 4J 4J 4440M 0
c.....40
0 6 .40040 04 44410o0.0--) 0C 41 00V04 W 04m
0 0-0UN10 000-V 0 41. 00 vi a 00W. 440 = a.
0.4
_..
r'
W.j 04 >0. V0 0. 0 00.
00t 0X
41
~ 01 a0U.
1o -o0Ul
0 - V 444W10. FaV041-V 0 0
4--
W0 0 V
40 04
E4404. fn
CW .. m 0 Nay 41> -4= 0W
C.C.:4 Q40 W 54
4 W.40 0~ Q, At00 00064 00 00 ..00U0 00
4)T040
k00. 0 c1 A 04W V0044
4:0
W oV4) E004EnU 3. ( 0 1 CL 0 0 0
0VC ?-0o
0 1 0 ;.
0 M 0 x0
4'0 %.44 0..00.44.
90 mI3Z4
m XJ.4 04 >1to
4. 0
tP 10U r0O -:1 0400 010W 0.4 4.4100a
00 -4-- a ''- 00>40 4) - 041 0W 0 010 (4041z
loc1. 0. .. .D -to 40 *&0.
-4 0 coa0441W1.
ON a>1 0 4? 0 c404 c
c0 a .ev0 =V4 01a - C4. 44C0Wm
k0.) 004. 0
ma4 r.0' 00 .0 -> 04W
4.) 4
040 440 4J 0 4.4 01 W 44
CO:4 4 ... 41
>WV.-.-2-
1%.-4. --c44 O~ .4Pr 4, 4j.
AC120 a4 V4. 1 4. 410-. x 0 000
04(0 f x 4 .4 ON
0 04 41 0 84
a0 l a.4144 14
oo&
0
N
4
C0-4 04 M
~
% 0.4.4 .)& 06. -4 C -4
- 01 1
r,00 4 C U 4Aj A i
N0 44 xa00 c
14 4 00m 60 MW4MLN WQ
0 4 1. 0 .. - v 0 AA 0
113 'A a' 01
z ~ - 0-
j)1
0 1 . - f=
a01 0d 0.0 4
-4 & > 1 11 1 C 1 4
W 4
CD44 (q wm
0. 4ff o0
- ,, 9>
4.p AJ
0.0c4-4.-'W fo
U.4 0 1 -4 C 40
B-41I
~0 301A0t
jZ
0 0-W W
VCpx
0
C
00to
A
' 0
4 0
. 0 'l43.
04CI 4.)A0
41 Z
0 4- N
C4j 9W4
0 v 40C
Cfl
0e
A
00'I
m 4.0
aM 4j0N-v w C0'3.. 0( 0Cw0 I A 0- 0 1CCU3; v 0
Tfic 4J &;Az
430r000M 0 x .-. 4 0 0Q0 00 000004
WA
Q
r v 0 r in-- vC0 0 23 >.'34
(0 >->0
O0. 1W a0all 4.4j 04. g C0 :X
0'344
00 ON 4 N 4.~.40 C4 (.. 00 COO > a
000040 0
0 E'0.wfl01.C -4- 0 a W0x* C O.0
to 4.' O
4 0Ud o N
*I'-
' >.0
'4C4 00 40'
w. 30 e 0
0. 0 0 Nz .1
3- 43 -c9 <0 3> 0 0 C .0>I 0 4I w r.4 r.- Wo a ONr A $ 004
ul 0 $0-4Q) .. 0C4 OC 6000Mm4T.0 M.. 4 4In 0 I 50.0 0 0
4.3
Q; a)440 a,3-
ECn' C 0rc0 45.a4.w. 04 (na, m o0 0 O.1 m 0
4jwz 0 N O > 0to0" a'0 0.0w'0 0 0 .4 00 0 w4304 - 0 0
0 0 0 , L r= rf 0- C O 0O m u - 4. a -x 4 v k 0C0 0 4 4 . 0Oi
0 4 w.A0 400
'
3t 4j 1 00 40
w. c3 -w . O4cO00w 0.-'>
0.4 000 l
0 0 A w 0 0 % 00 0 - w0 >.0304434w
0 0t
0 c1 waaZ"a- 0 04'443
w04 0-'4 C l CM.4
C-.
c0 tN >w 0.44 -'o0
co.OON' *- 0C (a0 >CO -04.04 04.j* a34
0004 C N
'-w'0Q 0 z
-0' L 1
00M>'a04 00C-
m C-4Z s z w iM
0 j 0O w
0'-4 -aC-u
a.,- m o - z 3 &.m ( .Vu.0= 43N3 3t 40 4.E :1 0.M- CZ 4w >>0dC w -- '--4
4)-wz.- w.''n z0-*0 u) w-.-.'N 4440P) .40 wx 0) AC0 04.0W-V .4
-0--D0, 00 C-f0.Cf -04 V 0- 00 C -40- 30 0'4
C'C00000'N ~- 2io '
>
0.
.'- .. o40.-'-
0f03. 0 0 0
cm r.C
09
4.4.44
0 N
OC
.04C
'0ll 0
00 - O 0 00' 4J . -0..4C*0 00->0 w3 w.4C-O4C
a
-4U0r
~40
N
o'40()
Co
-.. 0
mN
.C v
a
-N
C ws0
00
-- 104
a4.4 ~ 0 0.
a4.)44-wO
0oc
cC z .. DI-C4j
O w
M.-U
% D
0
3
w
04 2 >..0.'40
E- 04J0 a
0~0-'.-. - I O W 0 4 ..0') 0 4.0 0 C0'
O-. u01-C3>.4 Z
l 0C( -4QC 0 a-t 0 MO
.oN. -.-- 0C0 W-w 0-40400N
a
w.'.0 w m C AV0W r.0 4
0~ .'- ~ w oj 0 00-NV
N-C04.w. C u
tP
C04
to M
ME 4 U0.
0
9X04 000.4
V~ m0.0.--
a,<-~0 .4 0 0 w---
a, .00 >..C4
0 E- . 0 Z. '00--f 03- 00C
(0 >0 0R 0 E 10C00 *- a w V00'0.4
NO" X C3_N 4) G)4_ )ON 0 c .034 0 CI A u) w . -UCOr. MO C (C 0.
E4 V -v-~0. 0 m u) v . 0v-. r( 4 j0N3 ( C4) 430-- to. 04 0Q0 4 0 0
0Z3 2C0 4N > - '0,00- -'n.t .0 3z"N~ G 2OZ 01 00>.. -4w . 0 CO l0 0 u3
0 ' .0 4 ( 0a 0 00 . -. 4 .. N(a 0 > '4- f 4 0W 0 0m 0Q 3 0 0. M OC
430 u4 0> a)(00 03 = j3W 0a0( 0 > 0 0 .C0 00
NO <0 r> 0 N>00 4--C 3
W0 W 1%w 0- O 00%C4Jol 0 0 0 .I
U0 X0C Q) - L, W0N-' C- M.-W 4J4.~4' >0 000!000,j w b) 4 0 >0 0000
<.01 0>.- a 0 m-.-.0 0 *.. w4 No .1 00'>Ao. 0 0.,04 0 174j -w00
- '--O 0.0- z 0>o- -'.44 &n.000-- a :104 3 '-o -4.- r CM ErC0-. 1 >. 3 4
NQ4 o , Z. w-
ICN0430'4.l.4 -4V0 -
A0- WQ W.4.
ia 4 N 4 - N' 3 0) p> f
Ea W-4 0' 430..W
4)O (
00a -4- $ 0. W 0 0 0 4*3. 0 3rr.) L
-'40
'-00- W v) 4J
4-
0. 0 00
4
4.4 ~0 V4OCw1
m0. VMC
cO
1
L W4 0"' L, l4)
-'C 0 4 o0 4'C 20 r 8 43> 4
'v a) -- NO 0 00 4 0 -W
C W43 .O( 0 0 4.
0
43
to4--- 4j'C Cwm4. > 034. 0 r4.C40 <0C 4j1 .00
44 .C 4 00o3--au
. 0 430 a lwo, .0
-4 0 M 01 0 &U, 0a
000 C000<0
04- 44. Oj w0
0~~' - I 00 V 00 $1 Cw
0 3.- to v) 4 w-3 W0'-0 04.,u or V l z M?
,a
W 3 c -.
cCC
0 -C
0' . 000. 04.c
C14303a 0 4.
m
.
r
ul
00.0b
4j
0 1 ~ "4
w
C.003.
0 0 v40
4
4. c0 4.4. 01 0o 000
-u >.Vz0'4
L43 *Ca1I200.
0 0 04004 j u ,I r r>
1 w luICOn l4CVt 3. 4. 0W ''- 4 W aC)
S 4 w0 Z0a 1 l 0 vu 6 rwtaG60> 4.
0-'M U 44.4. z. w4 j 4)0.3.C
4 0-4.
D4 . .0
--- 8 . C u
>0a00
m- a
00c
'D
w 4) .0 UF V2r
A07'!-S
.0 o0
0 1
0')E
vV- A0
XJ.-0
jc*( a w
000-'4 r 40 C00 x3 s-
- 0443 4- 0.4J
044j ~ in 0 0
0 .0
4) 9 U~** . 4 4104 02 4. 0 4 .-
c 44
00.00~.0
CW 0 0N.- as' 44
m E-0 0 W.-4k 0 -'-4444V40'0034j
0'44 U)- 0 000A - 04 cU 4 0 *'04C X..
0W0A C
'O uIx0 0
CO/ E 4440to41040 0- 10 V3 0A0 0 W M4
z mm o 00.1 000
00w .. 4-4J H 00
IE441HVCCm4 D 0 (-go4
a444.V -.
a, , , 0 lp
, Z0, C 0 .CH
4 0 L0 w 0 4
c 09 '0 -0 u00-404. Om.4 0 m-0 N 0C 04..44-
wC '.f 040. 0-4
9.0 r-C '0 0~t 4C '. 0
Cr,
E-L,4
00
) 40'CL4E. M
*l
w
0 -
a -
< 4
m0C
t" 0. E-
co0 0 k 440-
4
900 90-4W'
0
00 0 4 N 0-~ 0 44 00 4 0 N 40-V
~ 4.
0 -0
00
v
;U
0.0
a o w~- C
00.4 0 Q
0
6WV
.- 4 .c i
1C.
0 a>r
0
0
to4,0 U . 1.2444'g 0 C 044 0
44 0am
4.I0_'0
ow "a20'
-
.
-.
41 w
0,00
4O40
m 0c :4
00K
~ . V C0.-0.44 Id044.
0D 104 0-
c ,-44 CC to
4c M1. 9 04 v4-.2 "4 0.-0 Z 0C.
r 004 -.
*-.4 CO4 >~~ 00 4.CC)~0-4 44 NCe0 .
u E00 O 00-. CO 0'. 0 0 0,2 .
C 000
0 0 -IOI 00.4
0o
44.44 0 1 ~4 M 0 0
410 ~0 V4
l
0-s 0-04) 4. . 404 0 4 0.
00
0 4 04 004 0
44
fu.-
0
.~40
: W
1~a-C0
0200
0. r4;464U . -. j0 1.2 0 or .
4
010 0-.0 CC yU 3 . n~~8
000 -IC rC
44440004n0$
g ;10U 0 -4.0
a C *W D
O'W4 4 0 C OU( a W44 QC.V WO V ' OW1
r.-C V 04A WU.40C.
0.Er.CW~4l w
o4 4) 41V44UWV
(U - (U CU) 4V-0-
V Qwo 0 4 00 (U(3 o V- ( O 300 0o a
a)01 *.- 0-r r0 >10 . >4 k0(Ja' Ho 1U *-0 0 .4-.a (W(U--a.ca >
010 c4 0%(-4W
0-4 to c W to Q..444' QNW (p- x *0 W -40'M - 0.l-
W 00 mV
Z4W - W 00 M - C - 40W k .4.0W oZ 4 U 4 - W 11 IA W 4'U.4444 --4o
C4>44 *4 C
0Z.kU0. C W40 C 0 (o C 0) 0 W-4(. Z WO. U(W -- ICC
a.(WU 0 N0 Z V.04040 kW440 0. CWCO :b oNCOW . c
-4 W'. c-4 rVr. W4 0.
(UZE-.0-4
0444 >I t y (A W 0. 4 W1
M(-I En r.>W MuO ,
(U.4( C0> .XI .0 C W4- 4 '. 0 3.-..
D4( V 4M'4' S4WQ
Nu~ w fl EUW V 00120
*.4VOW U $1(n N.44/ .) *-4 C(U4'A4C WQ w
r-c Q D
0,v
0x 00CEUIC~ 40C(
tnC>- OU C"
aM > 4-.n 0>.W 0 w0:
V 04( )
3414.0U''U( HC0(>(U-. N44 MCO>U- 4 'A U 4 4)4,c 4U
1% r.0 W'U.040J W40-C.~4 >U
443W0-402
$-4U
0)44 c Da4 $ Q )E W000 .C *> VCOC 44 l4 U
W-CU404
0
z .. =- 0 4(U-4 0 0 kOUOVCO4 . >00 a, r, uCCCO E 0-AXAm > W0(Cr
ar 0 1 0 0 0 444a,.C4W.,4xU) (U 00, a 0*4%0r= O0OW-40Uc
CD0t - 0UcUM 4'00 11 W oV CD 0 o- 04 U0- *431 cco mu0 . 0C
0 Av
CM >.0 0 t 0a0 Qv0m0 O) 000W
(U4'0u400W -U
> (UC . *VCW . r 4a, En ' 0 0 A .. 40 W> .
0-4 (1
E-
40
(U. -1
0 .0 mC 0- 0 X0
'.M 04
-W>N E- 4W(a -- V (UW09
( 44-
'04''N r..
A004E4C (U-WV4r 00 44)X ww fl(44 10 ZOC4(U-V
rA' C0.0> Q WW. n g- C r0 WUC V0 'A 4 'A 'M 140
4V40 -WC( oCVN 4V4 -03 AW00 -=0
z r-<w ) .14 - 0w 00z - - 0)r W 44 V40 01-Wo4 a
4 0-
r
4)0.
W
V *--4
>3Q.4
4440
o 1 (U ,
WV
w
v0x(
c 14 00 :3
>0..
.. 4 04402 >V
* 40
0 C*
Oa.r 0
C 0 40
(U4(UC44
-4 -
cC.W
-r a
0 0 4 -0.
W 3 -(U 4 4 R1 W4 X )aW- 0tP r.1 0C (UC 414 044 V :
UU SO2C (U 114 0' OO C( -
W. 00-)- W04 . arW00 .0 0 z c x-cV40Na)44
0 A> CW40 > V W 0 W0 p 4 D W-4 C.0Z 'S > 0 V0 W_0 =3
C)-,DV0.-4'- 0'W0 0 9O W0.04 V i , - .5 0) 1 4 -4VV
a --4 (U 1z
0 M 40 4.C W 0& UV)o W 0 '-40 0440 VI C 4) 1
0 ,W 0E4 W4 n . 2A 00O WC4 >
r.0- 4-ua 00 U3 2.4
.- WCCW.COOCwO 1 :30
-4U) 9,0 W'
' 3 o0 C 0-.- .W
-00( ,W *.4 Q
45a'Cn4 W0$
0.4
0C C OIc 0
0OoO,0 . C4 0 0 0 0 _ 04M u
C
w
.:j0
W 4-44
Hr
0044*
cc
U)F C4 -
0
$
1
r ~-0
WC
m9
0 U)W 40 W40 -40 1. wS0V
W 1 40
0) z
W
V~ to-
,
IQ~
1 414 ~ W)0
0 r44
NW 4)
44.E-4~
04 a v4 r jo0C4U
N E.4 m4.4- N4t r4. N~ E 44 00W .40.4
*a "W ( 0
W0(1 0 0 00u
V C cc34
4* wOC 04 4s -4 D
4* -04oV30.Q
r .mr c 0.-44*0 C 0
WO 'a .
44 v00 u -V40r
VC 4
0~.0 ~- '
010
'IC4(
11.-
im. 'an0
- NO .0
4 '
0 1C
0100 WW ' 40C3-
mnO0.4 Orl(
X0.4 U)rOI 4
Z4 0.o40
-40*.0 -,0(0 C00d -r00
C--E4~ 0 W*09
4 .O 00 0 E 04 :2.C4*W *0W4W
W -V )
4444. M4 0-1 a C 40
(n V MOMON V W-
44444 44-U IA)I40 v CQ.. 0 0 u
(10 0(4 o 0 -40'0-4% 0%O 0 4*o..
c 0(- .4.4E *O4'-
>44 14.S 0 4-0C a4 >ON
On00 40-U004
000n
-c 1 N-41 0 r
W r C0 0 I m). 04o4 0*
413Wto0 4Al k m In X
.e044I
t-' -. 4-
0 .4,0 t bU,4
r' .
00ol4 VW 4'oni 0, 0% A A
. 4 -. -4444*Z 4*
1A~W z W is4U
U ~ I4 ~FA-*0
& 0 - I 4 4 0*1s
B-44
c .
44>0 )00'4) 1WO % )Wt
M . aw 44-4
00 0 5424 G0 U l< ' 0 a,
M ' -0a12
Qo~ 0 w w .C 4 t04.43 4.000)
L, m ~ 0 ~ 0
C 0 . C 0 4)4. OCc J - ' 0 4
> V
C1 004)4 wC C. c.
4 54c
>44 s~ >~ -eo u co-D 0 c w a0
mcv 00
w0 -44. 0 0 oo (U 0 -4Q4
04.300--.04... 440'00~; 5.C VOC' 0 540 00.)
> z - v4I>300 .C4
w00- 01
. o%0-4 00-44)6
0 004.-4-
-3 C' 4.. .0 W0 4 0 05
1C43 1.NVC C' C W.
4C4~ a,)Z(0 00 440 wv
0 *..W4
C... 0- 4 -
*0
M. 0-4 0m. 4. >10.4
V,4 1) 00
04w. 0. j 4
W)0.C 4 00C - Z. 04 0 0
0W > W.C 00
r4 44 z'~. 00o,-
m Iv- 0
m-m 0o0M) 4.4.C 'a 0W A. to 4 -E0w4 .
o0
2 r r. 0C-M EC4 04
40Mmt Z44-4 0 r_ w0 v 0 *.n.0
*.) 4..0-. I 4)40
cW 4) 0 -4 ) .. 000 c4cnV 1.a 4 I 044 (a
En
:3 EO 0
W004 )0 U0-4:Ow -V 0 ' V..4 r-0 0 4J
W A u.
00 -4> CC 000.. Vw. 0
0042 0
00 . Q) 40 3 4) to M4-4,
0 W
144J-W
'$*-44
m 4 4410 V r
no4 r.:. C.a.U F4. w0
*- a w> >r-0 a0
4 0 0 0 * ..- 024.0
- 00 2 40 0 0 4J
w) *04-j 0 0 4 004 0 4.) M., (1W -4 14. 4J M:0r
<044V.P. (2.- 0 m4. C4 r<.4
04) w >. C 0.
4. u I o .
42 . '4 04 0w w4.Q x o~ f.a0
$.-40
r_ W
01:3~~
0 44
~ 3!ta47o
~~ ~ $ Z : ) 0:0 w a)
$4
a; M
a) u 004 oN a)
$4. 0ioo o 0 - >0V
=0 .V
;'-I o .z rC w044
5031' E C 0o
tr 0 0% a) 0
w.-W0r.Q0w.V 0 a - .- 0~4 H 0ww 04
> .. 40 - 04j
t3
. w 0.04).
Dw c = c- 0
m-0. 0 -4i 14w -~ W U .. ,.- m0 0
.410O v 4 L,.'4 0.- 4)
X.0-. 4 w .0.0
t m V 4.0 M 3l-
c
s. 4. w 'A(.4
0 ~= v ~~~~~~~~
044C
50
>(Nm)4) 3-4
4
> to a,
0 0 0- 0'0) 0' CC 0 0 . -440
l<- U 0- $4U. C 0 0 0 W 0 0 CM 43.4
o
4.= $ Mc 0 $4 0 D-00 Q. w. 444~ W V.C.
.: W04v r c0>4m
4J 0 04 5044 C040 < . I w .-
C: P .W a 0--4.40N 0 t0a
mCWi 0... U.
4--
,4"4 54
I 4t3)0Q oI40
C -.-. N 0 0- r0 0'.- 4.543a C0
40 4)4) to 0
z 4 w4-0-> 0. w 45z 044c 40030
0 .0 00 4 ) C0 W' 0 4j
M4 , 9.C w fC o.C 42 0-
m0 C 044 to 1-'4c00> 4..Z044004 0 1 4 0C4.4 P0--
V
3 4w0 _ - W >4)0-4.Ol
W =
. 0 .) - 00
V 0, 00 m 0 *.c
4-' -0't l-$ 050. 3 ODW m>-00..- 04) w2- W M04.- 0.Cs 4
tn MW000 C 401 0.O 00 4 I C - 1 oD 0 '0 o 0 440 e w >
W 0 0' r 3 M44 0 444 0 W40. 0nW4
3 c 0(Nr w)m)0 w.w4U44)0
M
4) t .00 06 0 V AJ 0- 4.04)U0m4 '40Zc- 4' 440U2 5c.- -0 I
.4405 N c 0VCOc
M V to o 44 mC a)0.<0 4j w E (a c 0
4 .. 0 N 'A-4C...,4 A w.. o 0 w- 04-
4.4
oc> 4400) 4J 0C 3 100-40.-4- v4-)44-.O 40
0 w C444Q0W0
-w m.U54-) 4o
CDW 0 0-OD0.-j 0W0 0 054 000(-
E*W DE 0Q I4 d)' 04J C 03 wV W-0 4C
:3 - ) a,40 - 0 0 W=a
0 00V =. a40 044-: - 4 0' 4J W - M N $4
4. . 1% (n- 4V 0C .0 05 0 0 E. at40 I04c)0wo m ..- a .
4 )..0
0*-~ ~~~~~~~
~ m..
.j .
0 I..
00.4 0 0-e 040
-. .. w 44 -4-w
o -4 .0( 20 w4) n 01 W r 0 a:C4
V - 0.
Q0 00 o4 w4 0V w4 toC~ M0C 3t44
V)>
Q) c 0 CC44) 002C 0 2 m~.4 l 0---4.0
.4C
01 ' 04 r; r4~ 0 .01-. Z CL>. C 0 0 0
17) 000000
(a M 00(0
0J ws 00 044.V)0
Itw '4=0
r_4j -- 4041 04)C.c0C ., 1(-O
4D4 0020 ~
.r ;a
(N (A40 44-40V 4)0M-4 r~ r m G0.CV04 M 6 24 m
.0 c 31 -4 to.4.CO0)
C0.5
E.4 4 00L
004 )
r
254
0
3t.
4)
0w5'405 4 4C4)
0W u 04)0w
0 W 40
PN
m. mC.
-,a)S 0.00
(n
0
40-4).
o G0()4
, o .0 ) ~
4 .u
w4-0-
0
14 m4).m a w w
50 0
.0
0 U
c
0 w w02 l 54V w... m.4 $-1 0 0 4444C
444 w
0)
oo0
~ ~ .
4
4-
4 j-0
~
4)
Wc04. w >-
.- CCa ~040 t ~ ~u CO ~ 0
<
:3 r. 40
.a
4)-.
0, w.>0
0 >1
0 n V 3 -W
0 0
w a>c j a
U04.- o- m(24
- c 4 - u 00y , mr w j0- 4A- 5
(N4
4'- 4.0 0 .0 42 5
0 v .0~
1. . 0 0C 0 0 tr
.44 4~4 1.20
0. 4) E-0. V WV 44wO.
.0>.>.>. M 0o 0 no0
.0-4 A)0
0 C-4
r-. 414
0. wW0 0 0 -'nV000a0
V00.. 4 o5~ 0- w- -- V.00-
0 w
0W0 r '0 "4'.44 0-4
4).~4 a. to w040404
2. 0n 0- w w U -- 0a
0 w' c r Dwa :1 n cc 00' 0 4)
4
st 0 0
0n4 0
to0
>04)4
M54-. Is
0 0
tn r
ll~~~0.
P r4 ar
w. W
m- 4 0 c .. 1 c w-w0 c >-%
11 r w mA4E
4)- $- a:54 0 to o- A0-40 > a :,; M.V c
m I 0- oo.
)- o 0 =)0 -04) wt j o r 0.
r 04 -341a% 0V'
0 -
.. 4) 2L U c .0-4 r-. - 1 1 0. a0 0 04 0 A
co a o w . 4)2.c 0 6V0 0)Vmc J 4 .0 4V00VV
a. o.5 co
0 i .
r45 0 00w
w4500C 'a
0
J.04
0
00.4.-
M0&Z 000H
:5j.
QCV0054
W c
....
.. 1 m v- 54NW4.m4
0
j w
00 m
VS>.-
1
0 0.5 0
0r
a,2
04 tom0 ~ c3 -4 J4 a4(010
$VS
4
. 0
04J4)-to
$ 0 Z .- V 060 'A 004
44 445.0V a, 04ix0w 4 00
4C
at~~. .0 c..4V.4.0: s j 44 0>
V>
m c4403Z I54>0 0 0 1'. oco4 00 **W-
J a
a<l-
0 0* 0, 0 'A a C 4a.03..400
-'a004 40A0VC J>54j c 0 X t4)
m" r to,V%.
Q VM0.v
r s.C :..L. M2 4 4' tno V0-0 ly M 4 .0 ..4 .2v
B-45
tr
0-- W ' 00Tr <00> '4IW000W -0 - r00 'a w
-0) r'0
le 40 0 > 0
00 414C 00-- 40
E0 W4) 00 1 44D
>4N CL 0 00 ( jZ 1
41 Q)
00404R404> 0.
1r v4 .-ONN 0 -0--4- 00a414/~ a4 0.4
'4 0
a ION---.j41 4 40- * W
-' 4 ) m4 44 4 0--
- 00 4 0 41 4)144IV4 0 4 CC- 1.w00ato0o0CU)00X
$ 140044140>
0 . i 0 - 4 > 0 4 0 1'0 44 03 V 440 0r 01 4 0 4 4
04 .0 ' - 0 1 4 -- 0 - 0 0
. - 4- 4 Q)40 A-- 0. 0 > -4 . 0 04 W 0 4 CC4040 4M 0
4J 0(0 000
-4 .- 4 00 4 004 "40 OZ : .4 4-)- 0.40.0 0 0 0
c0 1 0. 0;0. 0.44 a 0E>4 0 0
0 Y "407<E 0 C >W 0 L4 4-.4-
0 V V ( V-A M >20
OX4 0
v x %. 0 040~ ol OU) H 0 0 0 04N
0 000' a 09 0.0W E..W6 . 4 .0 C 000 a.0 to ?A
*-I. ).*C 0 4) 0' 0 0-- 0- 4) .1.0-
u01.*r J - C E. C. .4 04) -0 0V * 04 .00 0
0 I.V mM~ $)40 E0.00 0 W0w M o 04 01*4
V Ia
C W 4j~ 1' %.E44 40
tO 0 .. .O 41.W4 4J OD .4-
j.V
M co
.L 4- - -4 w4 . 04 a44.c-. 4
M 0 .00it
1400.- %4 4 .. .44 0.4 0 4 0
0 0003~~ 0_ 4t 4.04 0' E. 4v 4.0oI000
0c ms r C40 04) Q.~ V)4)4
Ca - 0p- V 00 0 00k-40Q
04r.c00 $ 0 4.
w0 4 00
V ,4
4j 4) c H 00.4 w
" H OI. w 0 w0>
to :3 0 044 114 31. aw0~ 04
.10 W V0?tCM jIV.0 040 V W4CLE-0 .
NU0) '
39-HV 0I -O 0>c
e &C 1_V z4 V0
9 x 00
MC.. 4) ;
Nv 4 , UNO ... 4 - 4 C N 0-- 0j m ,- C $10r:
E-
U) CIO 0-m-0 4) w . 0 3
4) " 0'1 VI t 01 0. .0V 0 44 10 . C
>... 0 41.. w> CM 0 4 0 * 43 4
4)-
A 3t 0O-4j O
>1= VV. UM0 4(. '.0144 0)0 4)e
0 14% C U 1 0 40004
to.
0) 0 VN 0 ) - 0 4 1$ 4 a.
000 E400
-C4J0c.) -.4 C
04j' 4.Ol 0 1~ 0 Ctu -0 0.
*0.'C C $44 0W4-440 -V E) V14-. ol w Vu j&. 4) -4 0
0 4 to .0 to-4 V0 4 01. 0 .
91 '00) 4 a CL.
4)44 w-44
W ) - Em
H.- a)04 Go 0 00 00,
4J8
V...-1 M VIV V 0.0 W(
-44.4N 0 0
r.--41-. 4J 0
r. 00
Zo )9 --040 rz V1 a 4)41 .Oi O0 9
344MCO .. 0 1 >.1.
-4 V0'.
rr10r - u0,a W~. 410.4 00 Id (a4t 0 00.. 0c Q ' 40 0 1.14)
N ~04..10
4)0)
0000 w0c4 0044
0.0.1)0000 r V
0 0 WH..
. 04 4)
W4)040
O0Nc4.VO'
0
0040 o'
(1) CL30
0'
j
0414)
) 0
0 04
0.
tn 040
00 a 0
4)
0 44
0 0
41. %4
V 0 O 0 0
V 04 0 0 1)Sa
rU) 0 c~
.0000 004 w 0 V..1 : c C)0a0 -j -. 10 0 *,wSIa0 00F
U4 - 4)
V 4
<ZDO *.)40 M 0V0 0 co1000Go4) 000k00a
0
V) w
0 10
WW C: OD0.4 0>4$A C. C.C :3 0 V.4 >4 0 04.14 00.S.1
V. -w04
-) 0 c4)0-.-V0 09 0. -4C1. 40J.. r_ $01.'.0
VV --
IW00 40C V
%40004 0 W *VZ . 0C%000.004
z .2V, 0-W4 .0W0 Q 0X.04C.- 0 24 ..-010 -01.w0 C1 V-0 41Q t
< N.OH a. 0 : 4 . 4-00-
>4 to) 4) NO= 4 't-C-U.-00C4 00>
c0 L cmrt 0nM
. i40 0a4j0
<.UQ)X
r tl Z
U r .4.0m.
w.-40V
<. 140
MC N >)
4441--4Ia) M 0
s0 0-4 0
Jr MO
VO)m ~
''. w _ 0. 4etV0 0 0c 0 C 0H 0 00 44 04
11 .0- 0, Ix W ()..4-0000 m. 0g4400
[n.40, - 0 0 0, 0 4)*.4) :1E)i - .4 V4 C. wI..04 o0 0
4 X:.. W . v 0 - 44 0)0-
0 0 00 1-0)
U.0
--M>00 c 0 j 0 A 0.0-1 w 1 C0 - S -- 000 N00
'<
V 444 C6 0 0 w4>A .o40to.. .0.H4CN) 40. 000 m V '-400 84
4) 0 00' 4)0w 0) )0 4jr M -V N0. C -004 j 0V
'-1 u*-'Vw
w' 004C4 0 o 00
: 0 .10
Q' -4 OO0 0 4W )0
0 0V 0
C 4 04 V 40
a;co4
3-.zco<
0
H
) 00
.) 00)
O
W
4D _e
* 40
M 0 4
WV10
0co1
C .
41
H
00
Q1 O 0
Go 14
4)2-
.0 400 l 4.4eU
c~ 4,c0
-0. 0' M.0 X40C --.1. - 0 4100 U 0 00
Iac4
0I r0u aCDi 0aL
ODi
dc 0TOV .01 004. N 4L 4.0-r
4 a 1 ).-40 4
040 Q.- -4 0QA040
B-47
-
OC
0' 0 O V40C4
.04
VCCO
0
W?
4
n V 2 .
00-toE
040.
~ 00
.0-
C 4
00
4 -0
00
PC1N .
C-4
CC-b
m 44c0r-U V00ma0
,h
0t1 AC FAV44441r
~ WucoE.j4
W - 0
E8
o0
0.
w>0
00(
04-
N ( 4- 00 Q $4 a 0 1j414-
W C 0V >.14 00 0-
414a C 01 0. 4
4CC- 90 jca 0 . 44 -- 00
0. 01 140 00 MA'- 14 wGo o 4 u4044VC04
'a w r4~. j0 " I -0-4- 0.01 0
0x P. P3 t, - 40 4C11- to, om .. 04 0g -401 0H0
0N 0 1 r0 4 0-04r04
2.1c .40a-0 0- Q .004-r0 (A 14 4 a00.4o
0 0.-L'. 4-h0 C10 0-
r-4 r. > ., -40N
rC v >014000
N 0 0J1ANc
00 - 000Z4 - >. 0 4)-.CCO
r- V Ai1CV44.r
4~ W * ;C 44 A0 A) 0 4) 31
00..-40>r
.4 X~ '0 40. V 0t 440
C
44 40 00. 0 0 WA 0 w44-'-10V4a
w~.
0
4) IS-b. 4t ro
0--- 00 04 00 C0 D 0
W >C 000c 0- 1 4 r- v W- ja V
OD 0) U 0V-. 4) 4 -0Aw V40 =I $0 4 MV >
04110
>,c 41 a o-
44-.0 I I>.1040 fA rm -IwO
-- .044J.1440 044; 0N 01
La > -- 0 14w
0%w(A
0 V, 44
4. 0 or-
014r .
0
14.0-
0-0 c H0.0 M% r
00e-40U
ji
nI W 2 x1.>u
o-~
oc~~~~~a
w
w~~~"
CID444-5
aava
0IcIQ
r4
o$
1
r4e4.IO
6O
-00 -40 a
. 0
.0-4-at
0
44- n 0.0-02
44
a0 0
I.,-. 44
14
a 140
19- 00.0 c 1-4 a0 41
1
0-40V04CM44a.Q
0 00 a
k44
ItV
[ $4 v0-
IOI ,.0
OZ*g0
4
0- ~ ] 40
%
p040c
-
-0
C
1. 201
N 1
44 4 0
0
a0
.'W
c 01
04%
4
W
144
0.
co A -0 4 A.U000
B-48
U) 4 42m 0 O.2J4 U .o . 41 a4 0C
to 00,402 I -4
V.-4V -4
*7 &4 go:.%4 414m4I 0 40 C-444.
14
0 0 m H 0" 0" M
4 42 04 0 0
04.V -0-44 so42C
04 - O
0 ~0. ~~~~~~
'0' 04 V4. . V2 r..-.
0' uJ4000 0 4 OC )0 0
0 2 >0C0 4--4
rk442
42
V 44 40 4-4 0 2 .0 C II
V -to 4 - O 0 W4C
3 0 2 0 0.44 0 0 0
2 M23
2 4. -. 0. 4 *- 4 V24 4a-.t-
0 0
00/C2- w .04 0 0 00 ' 4 .-M 0
C 44
0
442-
041 .C 0
4 z to20. to Oa) )4 to4. 0 0 f
'42..0 C104 2 4.042 0.4a,
.4 w2r40
410. 0 0 0W4C
O. 4. a
r-S. . m2 00.0 2 V. V04 )01 V2
0>4. 42o42cc 4 -V fl.- a 140 tr
0.40 s4 a 142
04 T 0 .4 V 0 0,4
is4 4.4 42 421 0 042go04 042 $C* 0, 4 4
m44 42 WCM0 4 44~o0 C >.4f 0 r H c -4.to
0 w 0 0., 314V 0 004. 04 .C 2.)> w w
--444 t 0
4w 41 N 0
H2' ."
1142 0 0
0444 En c2 .. '(24424
mt S 440 .1-42: 0 0 u0 W
r-4 0 F1 U--Q, CQ'44 2 0 ~ "4. w 2 V
41 00.0 4 . 410 41 4-0 . 040 .4.444r.
m4 >. 424 4) 024.
14420420 to42 4400 M 2.4. 4
0420- w" -. M a.
>C44 0 .0 42'--0 0% O 00
Q 0~
4442 w (a-- w. :1 40£D044 VO 42 OOCV
M442 0 0424-4 M.
4 3: Cr 0. 0..404a
$14
42-. -0 . 0 04 4 1 >'V00 ) M V 41 g 0'
4 faUCo. 0 4 -- 40 44 0420-4O QID . f4 0 14 C6
V0 %
0. W $14 0 U s00
m 42 Q'. 40 0.. . 00 .V 3 0 W4E..4
4 0042I
NC1 :)40 410.0 0 41 4.4
au 0 09 CD042 2 0P -- 4'22 4
r 1-i 4 4 V 41 02 M4-. E4V4 444 0 2 0
4 M 0 2H" 440 -4 442. --o'Zr44r
. -4 - (.04 rx 4 ) r_
CO44:3 9) O J4 > 4-441 V. co04
zg2412 0.4C- V'a
Cx.4
10
44.w0-1-M
42 w4.-4
.44 4 0 4X 04
.- 0r0,.22 .-.- to 0.1 I( C 0 >4
44 .- N4404 >24 O -4 044 4
a.14 CL 0042'a 00U -0 4.W42-
r (71 0. 0
1. w Z42 W 0 nO a) cc0"w
42. 00 o2.
~144-...0~42.-..C4 4>44. z
M-444- 'A
.0 to u 4V
4
C.0
4-) 42444
"4.40-----(n 2-x.W 0'04 k C() 4 -y2 rE0H------42c *- 00 4 444EM
4
:j 42
w2C4 V0 0 t0 .4.44 *.aC- wr: . 4220w 4 0 C0 4.403 0 C42 r
0.024
r.0
-420 44 0CW42 >04. 2 ---
E Z-~ 00CtoJ0 4
C3 0., 1,-042 CO..4.00
(a2 '04 C>024.-
0044 4.4-.4 q-0.V t z- w-.C'2.2..
Z)
r
0 " ~rz 4lWL
MD-n00
4-1> '
) 4
0.
- M
4-.VmIC0
$ 41--- V2
mC
0 <4X(a
E 44 -- C
34)4).W2C V-
4M 0.0 . M
42.f.) 04.QC
4' 2 w0 0M420 4. Z024 - C.L04 W
4,
4to V)0:3. - . 42M
0 I0. A 00z v 0344 4 4M O w C ~ ' 4 04
X Q 4J Vl42--.Q
2 'Q)o
4
I0r-4C
V
4J 44 0 m fu u ODM 42 w c N 0u .w'4L
0 -0orl
t 0. m -W o W N30- V4.I4 IW 0 440-4
(4I 4) z244 4 -C 0 V 420444..
10
- 4D IC.V
4. 0 >4 0 0 -'40
go .44
V . >1 44 m-
0 0 ,t a,0 C--N.. 4t_ V C 2 j C4 C .C OW.i
0 E 2.
w . MC '0 C M V 04.-..) - w G " Q ..
J V--4
400 0.4
'Co0 s4 > 42)C0 4.-
P4 C - 4 40 >- V'
444a2( (L4 £0 C>)0 0 w2 0 x2.'Cv0
4.4 -.-
C. 4203 W 4) 041=. 0).4 3 82 44 >0 C C r
. of.r 0.tn 0t 0 1 0m- f 42 4224 fla .) C 44 C
44U
-1 4' 42 0a0 z0.- 44- 42 440 4 4) 0)
4
4-
pi <
W
<-00
~ *-4
..
V4
00M.-
44..
a) a4
3
41
4) U
w
A0 42
0
0'
01 0
424
r_
424..
404 ..
c.24
I
4400 0> 42 V 44 t
0 0>X4.4
'n V
-u42- C0. z4 04
0 . 42444
> 0- MJ N2EC
440 .0>
13 v- 04 44420. 0 0 *-2 -4
42Z4 .) 44 422 V 042 00f 0 0.0 42 -- 4''C0
ZZ4'o 424 U . . 0--r" 4 4 V4.~4 44, 4
0.
CC. zZ MC V4 r4 >. a2C0
04 mN
C. N> 0c "- 04
Ww 0o 0 42 444 4 .4-
41 Q)44 a
C 4. 41-- 440 h 42. V 0 w 04 VOOMUc
0-0-
>' 3 -.44M
v 040440 aC).-
I.2.4
04C4 ' m 4) r4 0%r4 04J 00.4 C424o
. C C. C6 C 0 42> -0 04 0000>0X
'0044~
~~~ 44 Z'.--4
Cr
2- 4'-4
to
44440z CEn
0a 0
.02 002-
0 00 z" 4144 0
V3
04
002420 4
04 4 4 41 0440
CflIm 0 V3
-4Z. 2 4 o,
C.4 M Z c0
V4044 424 ' a N .5 0 ,2 2t 4j .,-
-004
o 444 .0
0 - - m 4 3 0 030-' 04u0 v
30
000 4 >( 4=842 4 v '.40444244 )
.CZO; 0 .. M24.. )-..'..4 0I2V0~
'4.V.41 W*i
W C $(4-0 44440 0444 C04 1- z~.42 0 -4
* a J440
0 T0 &14-4 4 .0
u- "4 Q 4 > 4 'a0.
0042 .kM 0 0
4 400 2 C OC 024 -4 44244
W24 M , 044.2 '.- o 0) 4 244M
>~ '
.4403,Q
02 z. 0 O .1
'004 42 0n09 Q
1-
0 1
'0"' u-' 6 Z>~ $ -. 0 c
Z44.' m44 V43 Z c .4 04.4 4
4 0.2 42 a E42424201
u ~ .4)Q3
044 44A
v0 cU __1__DH__
B-49
0 -4 OX In
0.0 C 0J r 4 -4-' -N C 0 90
a~ 0
ON 0 c 0
=0 0-4 " 4M4CM00
0 00 0.0~ 3- 0 o u-
Ai -i 0vm00a . 0
r
00 J mf 40001 U.0 ",
40 En 2.0 ~ 4 -.. 0
0Cl
E- I0. 0) iO0 uri2.z 4>
0 0 4.) tr 4) t c I w .40 0 a
> '0 4)~
0 40)0 4)a 0
0 'O0 0 V .0.
0Z a00
f,0 1 Ic w 00 4 . 'o0 *( 10 4
0 ,I .M 1 .4'.01 C 4 0v> c .1.-
to 4 O0
V.- A4J 0 VX* 0 ' 0 05 4
m N>
1_ rr 0'.- 10 4J 11'40--
e* 4)n
0 w C 15E,0O .110--VO* 4'
00 V IA r40 31 0 4 0.C
0.0 0 004J0.404J0.01..
C- 1
,.c m C,. 0 N '4 X C0 . 0 .0
40 r4. 0 - O 4 ( M W.4 0 4
V $. .- 0 *V.0u 0 - r2 0 to
-0 4 N04 S.'0) .0 x00.'A 4 0 0.
4) z
00' ). 00J0 r;0 .4.. 0 0 '
-0 0 U V x z .00 > Q0 go
2.c 0'40. 0 2f . w00 -4U
L40 Z ~ '0 o0 01O4
co 2.D 0'40'.0 430
~00 0'.,-
(N0 a,'0
L.0
0
00,
V 4) t 40) 4
k 4
0 .0.0
4) 41
0.41
-40 O0 04) 0-.00 IS .0-
ZOc .00 to 04 LO)- Q)-
'4
WOO 00. r -0 0000 0 01.00
co t *..A -0 4' ..- >0C 0 04)4j0' to0
0 00K0 41 2
'4O .. tr w
E- ) .U EnA trxOO
4-4 -. 0 >.0'0 t 0 W4
4.4A( 0- m 444000004a) JV
ZtJ40 *- o4)
0a 0 3 . 0to*
0 0- m44.4 0to-4 U m 0'4V
0 r0.0 -4 L 0. u s c On 0> )
4.1 0 OW M u0a00 0 .40 0t
>0 r0 w 0. w $41 >0-0
0I-*4) m->
0 c rf V m0c- r-0 >.1Sm'0 u' a f
'0n U-0 - )a 01C;a0c N.00 401 0.a
N~- ~~~-
00 . 00 4 *1 0-0
0o 3t1..--r
is2~
w4 00
-0.
W 0 M4
4Z ) W0-- -0 U W 10 0 . m
2)o0
(n 4)4 .04)0'zc
* 0 .1 Q-..u0 W)4J010
c0 4 0 0 *0 0.>-mO0 0~*...
P-0
8 -1 u 0i w3. 3 00HW 0a3.C-'
REPORT DOCUMENTATION PAGE
1. Recipient's Reference 2. Originator's Reference 3. Further Reference 4. Security Classification
of Document
AGARD-LS- 182 ISBN 92-835-0683-9 UNCLASSIFIED
7. Presented on 5th-6th October 1992 in Bad Neuenahr, Germany, 8th-9th October 1992
in Gebze-Kocaeli (near Istanbul), Turkey and 26th-27th October 1992
in Ottawa, Canada
8. Author(s)/Editor(s) 9. Date
Various August 1992 a
14. Abstract
The Lecture Series will cover the field of airborne and spaceborne SAR with respect to its
technical realisation in order to convey the participants' ideas and know-how on SAR, on its
capabilities and on the technology necessary for the successful construction and application of
airborne and spaceborne SAR systems.
The basic principles of SAR will be explained and SAR will be compared to airborne and
spaceborne radar with real aperture.
The influence of the antenna parameters on specification and capabilities of SAR and the
advantages, necessities and limits will be considered.
Digital SAR processing is indispensable for SAR. Theories and special algorithms will be given
along with basic processor configurations and different processing techniques on a hardware and
software basis.
The simulation of SAR-systems as well as SAR-products will also be a topic of the Lecture Series.
A presentation of the present state of the art, giving examples of presently planned and realised
airborne and spaceborne SAR with its foreseen applications will conclude the Lecture Series.
This Lecture Series, sponsored by the Avionics Panel of AGARD. has been implemented by the
Consultant and Exchange Programme.
,1. .. _ _ _
E' uu E , c
<0 <
AC
54 to
- CU .. .0
'a
-g !g .
j. ~2.
Cum
00 z 00
-u 00 M -2 -
<0 o .E -
ci < ua v 2 E r-
Wu cm >0U cm ' = u0
-0. 0 -1uI0
~CC
0-r r- cr
E- 1.I0
m E~
V3 C.,
13 -S~
~~ u0
-L do <~AU
0 -F -
5110 21
"Do.iiUo%.0m )
W0
6 2 . .0 .
0--i
-a IiiAl
03Q ,I-0
G..E
0-0av
Cc ?000
1~E
ID- -0 0 0
rv-A
NATO OTAN -
AGARI) does NO0T hold stocks of ACJARD publications at the above address for gehil dldtibutlon; Initial distribution or AGARD
publications is made to AGARD Member Nations t1rough the following National Dltrlbtlos Centres. Further copies are sometimes
avnlahle from these Centres (except in the United States), but if not may be purchased in Microfiche oir Photocopy form from the Sales
Agencies listed below.
NATIONALDIISTRIBUTICONrCENTRES
BELGIUM LUXEMBOURG
Coordonnateur AGARD - VSL See Belgium
Ftat-Major de la Force Afrlenne
Quartier Reine Elisabeth NETHERLANDS
Rue (l'Evere, 1140 Blruxelles Netherlands Delegation to AGARD
National Aernste I oinratory, NLR
CAN : -
LA.
GRE
ISBN 92-835-0683-9