——————
| deviation
degree
d deviation of
ig models in
of itself were
nodels of the
t meaningful,
1 be expected.
evaluated by
idge with the
' of the ridge
points of the
ted the ridge.
n of 0.4m and
n. The worse
| issue of the
s of the caves
accuracy, is
to be around
the standard
laser scanner
s data set and
The selection
houses. The
elong to one
The standard
ne is 3.61cm.
anes of larger
's. The main
considered as
—
viation [cm]
ses
ed plane and
wiss/Dresden
K
veloped, that
irborne laser
orrect results.
e mean point
orithm works
and 512MB
he figures of
increases, of
point cloud,
n.
Istanbul 2004
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B3. Istanbul 2004
The laser point cloud that the method requires can be obtained
in different ways. The laserscanner data can be segmented with
an image processing tool and the building polygons are than
used to select the appropriate points. If ground plans are
available, they can be used instead of the building polygons.
An operator can also select point clouds manually. Thus, the
method can be applied under various circumstances
Section 4 confirms that successfully reconstructed buildings
models can be an alternative to photogrammetric models
measured in normal aerial imagery. The accuracy in height is
superior to these photogrammetric models. The position
accuracy thought it depends on the point density, still has to be
improved.
Furthermore, the algorithm is quite sensitive to errors in the
laser scanner data. Poor data accuracy will prevent any result.
For optimal results strip information should be supplied with
the laser scanner data. If the strip information does not come
with the data and the strips have not been adjusted sufficiently,
the triangle structure will not represent the roof face properties;
the roof is not detected. However, the method can also process
rasterised data.
Figure 5-1. Example of building primitive reconstructed from
the 3D cluster analysis information
In further work this approach will be extended to also be able to
process flat roofs. This has not been yet possible because of the
error definition of the laser scanner data. Occasionally it
happens that only one roof face is detected and modelled. Still,
a tool has to be written that checks the laser point cloud if there
might be an opposite roof face.
Regarding the success of the method as a function of the mean
laser point distance, further analyses have to be invested
especially in the parameter space. Limits of the method such as
minimal possible laser point density and minimal laser scanner
accuracy that can be handled still have to be found. Within this
analysis the accuracy of the resulting models has to be
determined.
6 ACKNOWLEGMENTS
This work was partly funded by the Swiss Federal Office of
Topography. We thank the Swiss Federal Office of Topography
and the Dam Authority of Saxony (LTV Sachsen) for providing
the laser scanner data sets. The author acknowledges the
contribution of Ellen Schwalbe’s work within the project.
7 REFERENCES
Anderberg, M.R., 1973 “Cluster Analysis for Applications”
Probability and Mathematical Statistics, A Series of
Monographs and Textbooks, Academic Press, Inc. New York
Elaksher, A.; Bethel, J. 2002 “Reconstructing 3D buildings
from LIDAR data” IAPRS International Archives of
Photogrammetry and Remote Sensing and Spatial Information
Sciences Vol.34, Part 3A, pp.102-107
Gorte, B. 2002 “Segmentation of TIN-structured surface
models” Symposium on Geospatial Theory, Processing and
Applications; Working Group IV/6; Ottawa, Canada; July 8 - 12
Hofmann, A.D., Maas, H.-G., Streilein, A., 2002 “Knowledge-
Based Building Detection Based on Laser scanner Data and
Topographic Map Information” IAPRS International Archives
of Photogrammetry and Remote Sensing and Spatial
Information Sciences Vol.34, Part 3A, pp.169-174
Hofmann, A.D., Maas, H.-G., Streilein, A., 2003 “Derivation of
roof types by cluster analysis in parameter spaces of airborne
laser scanner point clouds” IAPRS International Archives of
Photogrammetry and Remote Sensing and Spatial Information
Sciences Vol.34, Part 3/W13, pp.112-117
Kaufman, L., Rousseeuw, J, 1990 “Finding groups in data: an
introduction to cluster analysis” Wiley series in probability and
mathematical statistics. Applied probability and statistics. John
Wiley & Sons, Inc.
Lee, L, Schenk, T. 2001 “Autonomous extraction of planar
surfaces from airborne laserscanning data" Proc. ASPRS annual
conference, St. Louis, MO, USA.
Maas, H.-G. 1999 “Closed solution for the determination of
parametric building models from invariant moments of airborne
laser scanner data" IAPRS International Archives of
Photogrammetry and Remote Sensing Vol. 32, Part 3-2W5, pp.
193-199
Rottensteiner, F., Briese, C. 2002 “A new method for building
extraction in urban areas from high-resolution LIDAR data”
IAPRS International Archives of Photogrammetry and Remote
Sensing and Spatial Information Sciences Vol.34, Part 3A,
pp.295-301
Shewchuk, J.R., 1996 “Triangle: Engineering a 2D Quality
Mesh Generator and Delaunay Triangulator” First Workshop on
Applied Computational Geometry (Philadelphia, Pennsylvania),
pp. 124-133, ACM
Vosselman, G., 1999 “Building reconstruction using planar
faces in very high density height data” IAPRS 'Automatic
Extraction of GIS Objects from Digital Imagery’, Munich, Vol.
32/3-2W5, 8.-10.9.99, pp. 87-92.
