relief areas for geocoding would give great flexibility 
for image fusion. 
It appears that there is no clear advantage in any 
particular filter but that attention needs to be paid to 
blunder detection. 
The sources of error in the stereo modelling process 
may be atrributed to a number of causes which 
include errors in azimuth timing and range. Such 
errors are not necessarily due to poor orbit 
determination or errors in the SAR processor but 
maybe due to the presentation and interpetation of the 
data in the header. Errors could also be due to weak 
geometry and measuring errors in the image. Future 
work will concentrate on making the optimum use of 
the header data available and improving the geometry 
of the viewing and tie point measurement. Roll-tilt 
data will be used and the model developed to include 
the provision for the use of more than two images. In 
the longer term other data such as SPOT could also be 
used in the same geometric model. 
ACKNOWLEDGEMENTS 
The data used in the work discussed in this paper was 
provided by ESA to University College London as 
principlal investigators for ERS-1 and as participants 
in the OEEPE/GeoSAR test of geocoding SAR data. 
The geocoding software was used with permission of 
GEC Marconi Research Centre who develped the 
software for the Defence Research Agency. The 
assistance of these three bodies, and of DLR, 
Oberpfaffenhofen is much appreciated. 
REFERENCES 
Chen Pu-Huai, 1993. Extraction of three dimensional 
data from European ERS-1 Synthetic Aperture 
radar imagery. MSc report, University of 
London. 82 pages. 
Clark C, 1991. Geocoding and stereoscopy of 
Synthetic Aperture Radar. PhD Thesis, 
University of London. 
Clochez O, 1993. Speckle noise reduction and 
stereomatching in ERS-1 SAR imagery. 
Training period report, UCL/ENSTBr Erasmus 
programme. 
Denos M, 1991. An automated approach to stereo 
matching Seasat imagery. Proceedings of 
British Machine Vision Conference, Glasgow, 
1991. Springer-Verlag 
Dowman I, Clark C and Denos, 1992. Three 
dimensional data from SAR images. Int Arch. 
of Photogrammetry and Remote Sensing, 
24(B4):425-427. 
442 
Dowman I, Laycock J, Whalley J, 1993a. Geocoding 
in the UK. SAR Geocoding:Data and Systems. 
Wichmann, pp 373-388. 
Dowman I, Upton M, de Knecht J and Davison C, 
1993b. Preliminary studies on the application 
of ERS-1 data to topographic mapping. Proc. of 
First ERS-1 Symposium, Cannes 1992. ESA 
SP-359:543-549 
Dowman I, Chen Pu-Huai, Clochez O, Saundercock 
G,1994. Heighting from stereoscopic ERS-1 
Data. Proc of 2nd ERS-1 Symposium, 
Hamburg, 1993. ESA SP361,pp 609-614. 
NASDA 1992, User's Guide for JERS-1 SAR Data 
Format, 1st edition. 
Leberl F, 1990. Radargrammetric Image processing. 
Artech House Inc., Norwood, USA. 
Renouard L and Perlant F, 1994. Geocoding SPOT 
products with ERS-1 Geometry. Proc of 2nd 
ERS-1 Symposium, Hamburg, 1993. ESA SP- 
361, pp653-658 
Shimada, M, 1993, User's Guide to NASDA's SAR 
Products, EarthObservation Center, National 
Space Development Agency of Japan, HE93014 
Rev.0..ex 
Winter R, Kosmann D, Schulx B-S, Sties M, 
Wiggenhagen M, 1993. Radarmap of Germany 
- first mosaic and classification. Proc. of First 
ERS-1 Symposium, Cannes 1992. ESA SP- 
359: 
Trai 
d'ok 
japc 
SAR 
Les d 
se so 
intér 
On a 
JERSx 
unive: 
les d 
altit 
carac 
le tr 
au po 
La dé 
proce 
préci: 
L'étal 
et de 
des c. 
décri 
stéré 
obten 
donné: 
La co 
des ‘€ 
SLÉLÉ