bul 2004
size at
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B3. Istanbul 2004
5 CONCLUSION
A unified framework for the fully automatic reconstruc-
tion of points and lines from multiple oriented images was
presented. Unification was achieved by operating in the
spatial domain from the earliest possible stage of process-
ing using statistical geometric properties of the extracted
features. The framework uses graphs induced by relational
geometric properties, that can be handled in a rigorous sta-
tistical manner. It was demonstrated using this framework,
that geometric information from the images is sufficient to
establish matches of points and line segments over mul-
tiple images, thus enabling an accurate scene reconstruc-
tion without radiometric information from the images in
the matching process at all. This indicates, that the known
orientation yields much more information than is used by
most other matching methods, that focus on the radiomet-
ric properties of the images and use the geometry only
to improve robustness and performance. All matching al-
gorithms, that are based on pairwise radiometric distance
measures can easily be integrated into the presented frame-
work. As a consequence an improvement of existing fea-
ture matching algorithms can be expected due to the exten-
sive and statistically rigorous use of the existing geometric
information and the possibility to integrate weak radiomet-
ric descriptors into the task of feature matching.
REFERENCES
Ausiello, G., Protasi, M., Marchetti-Spaccamela, A., Gam-
bosi, G., Crescenzi, P. and Kann, V., 1999, Complex-
ity and Approximation: Combinatorial Optimization Prob-
lems and Their Approximability Properties. Springer-
Verlag New York, Inc.
Baillard, C., Schmid, C., Zisserman, A. and Fitzgibbon,
A., 1999. Automatic line matching and 3d reconstruction
of buildings from multiple views. In: Proc. ISPRS Confer-
ence on Automatic Extraction of GIS Objects from Digital
Imagery.
Canny, J. F., 1986. A computational approach to edge de-
tection. IEEE Transaction on Pattern Analysis and Ma-
chine Intelligence 8(6), pp. 679—698.
C.G. Harris, M. S., 1988. A combined corner and edge
detector. In: Fourth Alvey Vision Conference, pp. 147—
151.
Faugeras, O., 1993. Three-Dimensional Computer Vision:
A Geometric Viewpoint. MIT Press.
Fórstner, W., 1994. A framework for low level feature ex-
traction. In: European Conference on Computer Vision,
pp. 383-394.
Forstner, W., B. A. and Heuel, S., 2000. Statistically test-
ing uncertain geometric relations. In: Proc. DAGM 2000,
Kiel, Germany, pp. 17-26.
Garey, M.R., J. D. S., 1979. Computers and Intractability
- A Guide to the Theory of NP-Completeness. Freeman.
1113
Heuel, S., 2001. Points, lines and planes and their optimal
estimation. In: Pattern Recognition, 23rd DAGM Sympo-
sium, LNCS, Springer, pp. 92-99.
Heuel, S. and Forstner, W., 2001. Matching, reconstructing
and grouping 3d lines from multiple views using uncertain
projective geometry. In: CVPR 01, IEEE.
Horn, B., 1986. Robot Vision. MIT Press.
Jung, F. and Paparoditis, N., 2003. Extracting 3d free-
form surface boundaries of man-made objects from mul-
tiple calibrated images: A robust, accurate and high re-
solving power edgel matching and chaining approach. In:
Proc. of the ISPRS Conf. Photogrammetric Image Analy-
sis, pp. 39-44.
Schmid, C. and Mohr, R., 1997. Local grayvalue invariants
for image retrieval. IEEE Transactions on Pattern Analysis
and Machine Intelligence 19(5), pp. 530-535.
Schmid, C. and Zisserman, A., 1997. Automatic line
matching across views. In: Proc. of the IEEE Conf. on
Computer Vision and Pattern Recognition.
Smith, S. M. and Brady, J. M., 1997. SUSAN — A new
approach to low level image processing. Int. Journal of
Computer Vision 23, pp. 45-78.
