d.
International Archives of the Photogrammetry, Remote Sensing
4. CONCLUSIONS
The main objective of our DRS investigations is optimisation
and homogenisation for relief shading of Mars image maps
derived from HRSC data. So far the presented results are the
outcome of our first yet most important development, viz.
fine-tuned shape from shading modelling of the Martian sur-
face DEM derived from elsewhere. Inspite of the still rudi-
mentary and simplified stage of our studies, the results con-
firm the considerable refinement potential of SFS for an ap-
proximate DEM. In particular, after future removal of the
main deficiencies in our approach, e.g., failure to comply
with shadow extraction, space-variant stochastic elevation
model, partial space-variant albedo, a further stabilization
and improvement for the DEM may be expected.
The refined DEM can then be exposed to artificial illumina-
tion for the purpose of obtaining uniform relief shading of
mosaicked ortho-images. For small scale MOC imagery this
may already be done now. E.g., the MOC scene in Fig. 2 is
shown again in Fig. 4, viz. as scene radiance derived from the
SFS-refined DEM (Fig. 2e) with different illumination direc-
tions. Fig. 4a represents the surface under illumination from
south-west, Fig. 4b from north-east. Note the relief inversion
in Fig. 4a, an effect that has to be avoided by all means (op-
timisation of relief shading).
Fig. 4b
Figure 4. Artificial relief shading of MOC scene
ACKNOWLEDGEMENTS
This work was supported by the German Ministry for Educa
tion and Research (BMBF) under Contract 500M0102 via
the German Aerospace Center (DLR) Project Administration.
The authors are very grateful to the Co-Investigator Team
members Marita Wählisch and Frank Scholten of the DLR
Institute for Planetary Research both for their frequent advice
and unselfish help in producing and preparing suitable data
for our research work. Special thanks go to our students
Bernd Jeschke and Jórg Renter who have made small yet im-
portant contributions to the software development. We are
particularly indebted to Alexander Kerimov of the Troitzk In-
stitute for Earth Magnetism and lonosphere of the Russian
Academy of Sciences (IZMIRAN) for his advice and help in
matters of optimisation theory.
and Spatial Information Sciences, Vol XXXV, Part B4. Istanbul 2004
1303
S. REFERENCES
Beckermann, B. and A.B.J. Kuijlaars, 2001. Superlinear Con-
vergence of Conjugate Gradients. SIAM J. Numer. Anal,
Vol. 39, No. 1, pp. 300-239.
Crouzil, A., X. Descombes, and J.-D. Durou, 2003. A Mul-
tiresolution Approach for Shape from Shading Coupling De-
terministic and Stochastic Optimization. /EEE Trans.
P.AMI Vol. 25, No.11, pp. 1416-1421.
Dorrer, E. and X. Zhou, 1998. Towards Optimal Relief Rep-
resentation from Mars Imagery by Combination of DEM and
Shape from Shading. Int. Arch. Phot. & Rem. Sensing, 32 (4),
pp. 156-161.
Hashemi, L., A. Azizi, M.H. Hashemi, 2002. Implementation
of a Single Photo Shape From Shading Method For the Auto-
matic STM Generation. /nt. Soc. of Phot., Rem.Sens. and
Spatial Inf. Sciences, 34 (3B), pp. 71—73.
Horn, B.K.P, 1970. “Shape from shading: A method for ob-
taining the shape of a smooth opaque object from one view“,
Ph.D. dissertation, Dept. of Electrical Eng., MIT, Cambridge,
MA.
Rajabi, M.A. and J.R.S. Blais, 2003. Optimization of DTM
Interpolation Using Shape from Shading with Single Satellite
Imagery. The Journal of Supercomputing, Vol. 28, pp. 193—
213.
Shewchuck, J.R., 1994. An Introduction to the Conjugate
Gradient Method without the Agonizing Pain. August, 54 pp.
http://www.cs.cmu.edu/quake-papers/... .ps.
Zhang, R., et al, 1999. Shape from Shading: A Survey.
IEEE Trans. P.A.M.L. Vol. 21, No.8, pp. 690—706.
