Figure 3: Aerial image
formula, the computational effort depends only from the size
of the model and is independent from the size of the seg-
mented region.
The homogeneity predicate is used in a region growing
scheme, but it can also be used in other control algorithms for
image segmentation or clustering. Experiments with synthe-
tical images have shown, that the most important parameter
of our approach is the variance of the noise in the image. For
segmenting aerial images, this variance is estimated using an
algorithm from the literature. Initial seed regions and the
model type to use is extracted from map data. The segmen-
tation results are good for non-textured areas and for areas
with regular texture. For irregular textured surfaces experi-
ments with higher order MRF-models will be performed.
ACKNOWLEDGMENT
This work is funded by the Deutsche Forschungsgemeinschaft
(DFG).
References
Besl, P. (1988). Surfaces in range image understanding.
Springer, New York.
Brügelmann, R. and Fórstner, W. (1992). Noise estima-
tion for color edge extraction. In Fôrstner, W. and
Ruwiedel, S., editors, Robust computer vision, pages
90-106. Wichmann, Karlsruhe.
Cohen, F. and Fan, Z. (1992). Maximum likelihood unsuper-
vised textured image segmentation. Computer Vision,
Graphics and Image Processing, 54:239-251.
Haralick, R. M. and Shapiro, L. G. (1985). Survey, image
segmentation techniques. Computer Vision, Graphics
and Image Processing, 29:100-132.
Herlin, l., Bereziat, D., Giraudon, G., Nguyen, C., and Graf-
figne, C. (1994). Segmentation of echocardiographic
668
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B3. Vienna 1996
Figure 4: Segmentation result
images with Markov random fields. In Eklundh, J.-O.,
editor, Computer Vision — ECCV '94, pages 201-206,
Berlin. Springer.
Landes, S. (1995). Entwicklung eines Fláchenwachstumsver-
fahrens zur Segmentierung von Luftbildern. Master's
thesis, IPF, Universitat Karlsruhe.
LaValle, S. and Hutchinson, S. (1995). A Bayesian segmen-
tation methodology for parametric image models. /EEE
Transactions on Pattern Analysis and Machine Intelli-
gence, 17:211-218.
Pal, N. R. and Pal, S. K. (1993). A Review on Image Segmen-
tation Techniques. Pattern Recognition, 26(9):1277—
1294.
Pavlidis, T. (1977). Structural Pattern Recognition. Springer,
Berlin.
Pearl, J. (1986). Fusion, propagation and structuring in belief
networks. Artificial Intelligence, 29:241-288.
Quint, F. (1994). Bildsegmentierung mit einem Bayesschen
Ansatz. Technischer Bericht IPF-FQ-9/94, IPF, Univer-
sitat Karlsruhe.
Quint, F. and Sties, M. (1995). Map-based semantic mo-
deling for the extraction of objects from aerial images.
In Grün, A., Kübler, O., and Agouris, P., editors, Auto-
matic Extraction of Man-Made Objects from Aerial and
Space Images, pages 307-316. Birkhäuser, Basel.
Silverman, J. and Cooper, D. (1988). Bayesian clustering for
unsupervised estimation of surface and texture models.
IEEE Transactions on Pattern Analysis and Machine In-
telligence, 10:482-495.
KEY
ABS
The |
struci
used
betwe
are u:
KUR
Das €
struki
verwe
für di
Metri
vorge
Unde
ing ta
based
eral y
Gable
Girau
Orien
1995)
are in
ment
Onec
large
fine tl
our Sy
doma
desigr
the m
of the
the re
respec
which
knowl
and ir
semar
by the
The
of the
model
autom
in thr
