116 
a 
REFERENCES 
Sympos 
Pr = P 7' 
G\e) л 2 
(4тг) 3 R a 
with: P R received pcwer; P transmitted pcwer; G(9) 
antennagain function; 9 depression or grazing angle; 
R distance to target (slant range) in m; C* radar 
cross-section in m ; X the radar wavelength in m. 
P and X are constant. The range R to a target is 
measured directly by the system (the time between 
the transmission of a pulse and its reception). A 
special receiver and amplifier make that the 
output voltage of the system varies linearly with 
the logarithms of the received pcwer P R . The antenna 
gain function G(Q) is kncwn iron measurements; these 
measurements were made on the antenna alone as well 
as when mounted under the aircraft. In the last case 
this was done by using test fields with a kncwn 
backscatter.. 
Now we arg able to measure O’ the radar cross 
section in nr . What we need, however, is the 
backscatter coefficient y : the radar cross-section 
per m . For a beamfilling target we then must divide 
the radar cross-section by the cross-section of the 
illuminating beam at the place of the target: 
y =<r/ A (A= cross-section of illuminating antenna 
beam). With pixel values given as the imagery 
will look better since the angular dependence of y 
of most natural targets is relatively small for the 
angular coverage of a STAR. Substitution in the 
radar equation new gives: 
- El {4n)} Ri C ° S 6 
7 ~ Pt G z (6) \ l PL sin 6 
where (l is the antenna beairwidth and L the 
pulselength_ In this formula alone the antenna gain 
function G(0) is dependent on aircraft attitude. 
This attitude is measured together with the place of 
the aircraft by an inertial navigation system. (INS: 
LTN 58), which data are recorded simultaneously with 
the radar data. The height of the aircraft is 
measured also by the radar system: the first echo is 
that of the ground immediately under the aircraft. 
A special algorithms (the PARES algorithms) was 
developed which takes all these data together and 
finally gives the data measured by the system as 
radiometrically and geometrically correct pixels. 
For a more complete description the reader is 
referred to Hoogebocm (1983). and Hoogeboam et al 
(1984), 
To make the pixel values absolute a reference 
signal is used. This reference signal is fed into 
the system using the free time (without echo's) 
between the transmission of a pulse and the 
reception of the first echo of the ground beneath the 
aircraft. On top of the above procedures the system 
is controlled at regular intervals by mounting 
corner reflectors with accurately kncwn radar cross- 
sections in the test areas. Frcm such experiments we 
know that the absolute accuracy of a pixel is in the 
order of a 1 to 2 dB and the relative accuracy in 
the order of 0.3 to 0.4 dB. 
4. CONCLUSIONS AND ACKNOWLEDGEMENTS 
Churchill, P and A. Wright, Human and automatic 
interpretation of radar images of land cover. Proc. 
EARSeL Workshop 'Microwave Remote sensing applied 
to vegetation' Amsterdam 10 - 12 Dec. 1984; ESA 
publication SP-227, Jan.1985, p.131 - 140 
Hoogebocm, P. 1983. Preprocessing of side-looking 
airborne data. Int.J.Remote Sensing 4: 631 - 637 
Hoogeboom, P. 1986. Identifying agricultural crops 
in radar images. This Proceedings 
Hoogeboom, P., P. Binnenkade and L.M.M. Veugen 1984. 
An algorithms for radiometric and geometric 
correction of digital SLAR data. IEEE Trans. 
Geosci. and RS GE-22: 570 - 576 
de Loor, G.P. 1981. The observation of tidal patterns 
currents and bathymetry with SLAR imagery of the sea. 
IEEE J Oceanic Eng. OE-6: 124 - 129 
de Loor, G.P. and P. Hoogebocm 1982. Radar 
backscatter measurements frcm platform Noordwijk 
in the North Sea. IEEE J Oceanic Eng. OE-7: 15-20 
de loor, G.P., P. Hoogebocm and E.P.W. Attema 1982. 
The Dutch ROVE program. IEEE Trans. Geosci. and 
RS GE-20: 3-11 
Moore, R.K. 1979. SLAR image interpretability - 
trade-offs between picture element dimensions and 
non—coherent averaging. IEEE Trans. Aerospace and 
Electron. Syst. AES-15: 697 - 708 
a. 
An absolute digital SLAR is new available in the 
Netherlands. It delivers images with pixel values 
given in dB. The residual speckle in natural targets 
must be taken into account in the interpretation 
procedures. The development of this radar system was 
the carbined effort of the following institutes: the 
Physics and Electronics Laboratory TNO, The Hague; 
the National Aerospace Laboratory NLR, Amsterdam; 
the Delft University of Technology and the Survey 
Department of Rijkswaterstaat, Delft. 
b. 
Fig.l. Radar images of the same agricultural area 
with different numbers of independent observations 
("looks") in a pixel, a. With 2 looks; b. with 30 
looks. Images, courtesy NLR. 
Devek 
G.Domik 
VEXCEL C 
J.Raggan 
Research C 
1. INTRODl 
Post processin 
usually meanir 
kinematic natu 
Thus the rada 
complex intera 
tones of the in 
be necessary tc 
One may think 
e.g. map data, 
of the image fi 
give a general 
an illuminated 
simulation. S 
can be combii 
After rectificat 
with the image 
Essential tools 
overlapping SA 
integration of t 
rectification of 
as archiving an 
programs 
and 
essential SAR ( 
2. WHICH 1 
Software 
can 1 
There is of cot 
(a) 
Gen 
The non-radar 
output data and 
not be available 
(b) 
Arc! 
(c) 
DEIS 
The radar 
softw 
(d) 
Groi 
(e) 
Sing 
(0 
Stere 
(g) 
Stere 
(h) 
Rada 
(i) 
Stere