cover surface areas of poor undulations with the same point
density.
5 CONCLUSION
The paper presents the developments of the edge preserving
filter for the automatic DEM generation. The use of the adapt-
ive information preserving filter is proved to be a suited
method for the automatic DEM generation and the detection of
breaklines. The concept for the processing of the results of the
surface reconstruction (e.g. approximate breakline location and
corresponding direction) is outlined for the automatic break-
line detection.
It is a well-known fact, that linear filtering leads to a loss of
information due to the global smoothing, regardless of struc-
tures in the data. Therefore we passed over to information pre-
serving filter methods. At first we investigated a global edge
preserving filter without an adaptive property. This application
has the disadvantage that the filter process is not in general in-
formation preserving, because only the discontinuities in X-
and Y-direction are exactly recognized.
In order to overcome this drawback, we propose to use an
adaptive information preserving filter. The implemented algo-
rithm in MATCH-T utilizes the local structure directions of
the surface characteristic features for the optimal detection of
discontinuities. The dependency of the rotation of the local
curvature maximum against the object coordinate system for
the edge detection is reduced to a minimum.
The adaptive surface reconstruction process provides the geo-
metrical approximation and the corresponding local direction
of the surface specific structure lines. The practical results
show the efficiency of this step referring to an open coal-mine
excavation area and mountainous terrain. We can see that the
breaklines are represented by a band of breakline points to the
left and right of the real breaklines. This fact lead to ambi-
guities during the attachment of the point information to the
breaklines. For the surface reconstruction process it is suffi-
cient to have this point information of the breaklines.
The optimized detection of the breakline areas allows us to
minimize the model error caused by the spacing of the DEM
posts. In local surface areas of maximum curvature we adapt
the grid width according to the detected discontinuities.
The concept for the automatic breakline detection outlines the
treatment of the point information of the breaklines. It was
shown, that the introduction of breakline information from
previous pyramid levels into the next leads to an improvement
of the quality of the DEM.
6 REFERENCES
Ackermann F., Krzystek P., 1991. MATCH-T : Automatic
Mensuration of Digital Elevation Models. Proceedings of
Technical Seminar of the Sociedad Espanola de Cartografia
Fotogrametria y Teledeteccion. Barcelona, April 12, 1991, pp.
67-73.
Ebner H., Reif P., 1984. Experiences with height interpolation
by finite elements. Photogrammetric Engineering and Remote
Sensing, pp.177-182.
952
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B3. Vienna 1996
Förstner W., 1989. Tutorial for Robust Statistical Methods for
Computer Vision. San Diego.
Huber P. J., 1981. Robust Statistics. Wiley, New York.
Krzystek P., 1995. Generation of Digital Elevation Models.
Second Course in Digital Photogrammetry, Bonn, F ebruary 6 -
10, 1995, Chapter 7.
Weidner U., 1994. Information-Preserving Surface Restoration
and Feature Extraction for Digital Elevation Models. Spatial
Information from Digital Photogrammetry and Computer Vi-
sion, ISPRS Commission III Symposium, Munich, September
5 - 9, 1994, pp. 908 -915.
Wild D., 1990. Untersuchungen zur informationserhaltenden
Prádiktion und Filterung von Fláchen. Diplomarbeit am Insti-
tut für Photogrammetric der Universitit Stuttgart.
Unveróffentlicht
Mo eo UU Uy OY NO
CS ^ GO D v") ial]
