Figure 4: Perspective view of Loess plateau's parallaxes
7 CONCLUSIONS
Array Relaxation for Multiple-point Least Squares Match-
ing (ARMLSM) has the great potential in the practical pho-
togrammetric systems due to its very high computational ef-
ficiency comparing with the traditional MLSM. The new for-
mulation of ARMLSM under the variable weight model pre-
sented by this paper makes it possible to capture the more
detailed terrains through the variable weight models which
are adjustable and followed the expected rule: giving tighter
constraint weights in flat open areas or poor image inten-
sity areas and looser constraint weights in rapid changed or
rich information areas. The preliminary experimental results
prove its ability to produce high-efficiency and high-quality
image matching on the general purpose computer, and the
results also show that very dense elevation models computed
by our ARMLSM provide the subtle terrain details. Since the
variable constraint weights can be applied in our ARMLSM, it
opens a new way for using efficient array algebra technique to
consider the breaklines, occlusions and discontinuity problems
during the matching processing. The strategy of choosing the
constraint weight discussed here is a very simple way, to find
a better solution for constraint weight models becomes our
current focus for the further developing of our ARMLSM.
8 ACKNOWLEDGEMENTS
| would like thank my colleague Maurice Friend for his reviews
of this paper.
REFERENCES
[1] F. Ackermann, "Digital image correaltion: Performance
and potential application in photogrammetry,” Pho-
togrammetric Record, 1984.
[2] H. H. Baker, "Multiple-image computer vision," Pro-
ceedings of Photogrammetric week, pages 7-19. Insti-
tute for Photogrammetry, Stuttgart University, 1991.
[3] S. T. Barnard, “Stereo matching by hierarchical micro-
canonical annealing," Technical Report 14, SRI - Inter-
national Technical Note., February 1986.
[4] H. Ebner and C. Heipke, "Integration of digital image
matching and object surface reconstructure," Proceed-
ings of 16th ISPRS, volume 27, pages 534-545. ISPRS,
Comm. lll.
982
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B3. Vienna 1996
U. V. Helava, "Object space least squares correlation,"
Proceedings of ISPRS, volume 27B3, pages 321-331,
Kyoto, 1988. ISPRS, Comm.lll.
M. Li. Hierarchical Multi-point Matching with Simulta-
neous Detection and Location of Breaklines. PhD thesis,
Department of Photogrammetry, The Royal Institute of
Technolgy, Stockholm, 1989.
M. Li, "Hierarchical multipoint matching," Photogram-
metric Engineering and Remote Sensing, (8):1039-1048,
1991.
R. March, "Computation of stereo disparity using regu-
larzation," Pattern Recognition Letters, 1988.
U. A. Rauhala, "A revire of array algebra," Proceed-
ings of ISPRS, Helsinki, 1976. Comm. lll, International
Society for Photogrammetry.
U. A. Rauhala, "Array algebra as general base of fast
transforms," Comm. lll Image Processing Symposium,
Graz, 1977. International Society for Photogrammetry.
U. A. Rauhala, "Array estimation in signal process-
ing,” Photogrammetric Engineering and Remote Sens-
ing, 48(9):1437-1444, 1982.
U. A. Rauhala, "Compiler position of array algebra tech-
nology," Proceedings of ISPRS, Comm.lll Symposium,
volume 26, pages 173-198, Rovaniemi, Finland, 1986.
ISPRS.
U. A. Rauhala, "Compiler positioning system: An array
algebra formulation of digital photogrammetry,” Pho-
togrammetric Engineering and Remote Sensing, (3),
1989.
U. A. Rauhala, D. Davis, and K. Baker, "Automated
dtm validation and progressive sampling algorithm of
finite element array relaxation," Photogrammetric Engi-
neering and Remote Sensing, (4), 1989.
D. Rosenholm, "Accuracy improvement of digital
matching for evaluation of digital terrain models,"
Proceedings of ISPRS, volume 26, pages 573-587,
Rovaniemi, 1986. ISPRS, Comm.Ill.
D. Rosenholm, "Multi-point matching using the least
squares technique for evaluation of three-dimensional
models," Photogrammetric Engineering and Remote
Sensing, VIII(6), 1987.
D. Rosenholm, "Multi-point matching along vertical
lines in spot images,” International Journal of Remote
Sensing, 9(10), 1988.
D. Terzopoulos, “Regularization of inverse visual prob-
lems involving discontinuity,” IEEE PAMI, 8:413-424,
1986.
M. Weisensee, "Modelle und algorithmen fur das
facetten - stereosehen," Technical Report 374,
Deutsche Geodatische Kommission, Munich, 1991.
B. Wrobel, “Facets stereo vision (fast vision) — a new
approach to computer vision and to digital photogram-
metry," Proceedings of Intercommission Conference of
ISPRS on Fast Processing of Photogrammetric Data,
Interlaken, 1987. ISPRS.
Y. Zhang. Stochastic Stereo Matching with Neural Net-
work. PhD thesis, Zhenzhou Technical University of Sur-
veying and Mapping, Zhenzhou, China, 1990. in chinese.
IM - PFN 08 MP N
Peng ppd ppg
es No €& FF
—
pu at Perk Y - o gue Sem pmb ey u
a ai 2