3. Istanbul 2004 
  
fan aerial image 
after smoothing 
(f) The extracted 
10del. Execution 
nd compared for 
an Intel Pentium 
00 of normalized 
with initial data. 
that convergence 
3 when initializ- 
contained in the 
id without loss of 
'M parameters 8 
initial segmenta- 
DTM estimation 
s during iteration 
1atic real time fil- 
dels is presented. 
evel points froma 
v large variability 
The required user 
/e the supervision 
ind surface points 
ametric DTM sur- 
tation time which 
eprocessing; and, 
M. These observa- 
y generated DEM 
ne laser scanning 
ery. 
re, H., 1997. Ex- 
| areas in a perl 
  
  
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B3. Istanbul 2004 
  
Synthetic dataset 
with initial DTM without initial DTM 
  
  
  
  
  
N = 1 N=2 N=] N 9 
Can c 4 Can SAN Char FW 852: 12 
100 1718 213 4794 
182 it. 268 it. 393 it. 753 it. 
  
  
Hoengg stereo correlation dataset 
with initial DTM without initial DTM 
  
  
  
N N 9 NE N —2 
Gnos 74 | Chan 94d | enar 20 1 enar £220) 
100 1153 238 3217 
54 it. 58 it. 128 it. 135 it. 
  
  
  
  
Amiens airborne laser dataset 
with initial DTM without initial DTM 
  
  
  
N= 1 NU A N = 
Cnar à 1 Gnax > 5 | man 20 l| 6522 20 
100 1602 322 5774 
23 it. 38 it. 91 it. 139 it. 
  
  
  
  
  
  
Table 1: Comparison of computation times and number of itera- 
tions for the DTM estimation initialized with or without ground 
level segmentation. 
  
Figure 4: The DTM surface model with N = 2 computed from the 
DEM points included in the ground level segment of Figure 3 (f). 
Relative coloring was applied for visualization. 
urban context, in : Leberl, F. , Kalliany, R., and Grüber, M. (eds.), 
1997. Mapping Buildings, Roads and other Man-Made Struc- 
tures from Images, Schnftenreihe der Ósterreichischen Com- 
puter Gesellschaft, Band 92, R. Oldenbourg, Wien / München, 
pp. 159—174. 
Baltsavias, E.P., Grün, A., and Van Gool, L., 2001. Automatic 
Extraction of Man-Made Objects from Aerial and Space Images 
(lll), Balkema Publishers, Lisse / Abingdon / Exton / Tokyo. 
Belli, T., Cord, M., and Jordan, M., 2000. Colour contribution 
for stereo image matching, Proc. International Conference on 
Colour in Graphics and Image Processing (CGIP 2000), Saint- 
Etienne, France, pp. 317—322. 
Belli, T., Cord, M., and Jordan, M., 2001. 3D data reconstruc- 
tion and modeling for urban scene analysis, in (Baltsavias ef al., 
2001), pp. 125 — 134. 
Collins, R.T., Hanson, A.R., Riseman M.R., and Schultz, H., 
1995. Automatic extraction of buildings and terrain from aerial 
Images, in (Grün ef al., 1995), pp. 169—178. 
Cord, M., Jordan, M., and Cocquerez, J.-P., 2001. Accurate build- 
Ing structure recovery from high resolution aerial images, Com- 
puter Vision and Image Understanding, 82 (2), pp. 138 — 173. 
139 
    
Dataset Zürich Hoengg, 2001. Eidgenössische Technis- 
che Hochschule, Zürich, Switzerland, Project Amobe II. 
http://www.igp.ethz.ch/p02/research/AMOBE/dataset.html. 
Grün, A., Baltsavias, E.P., and Henricsson, O., 1997. Automatic 
Extraction of Man-Made Objects from Aerial and Space Images 
(II), Birkháuser-Verlag, Basel / Boston / Berlin. 
Grün, A, Kübler, O., and P. Agouris, P., 1995. Automatic Ex- 
traction of Man-Made Objects from Aerial and Space Images, 
Birkhäuser-Verlag, Basel / Boston / Berlin. 
Haala, N., Brenner, C., and Anders, K.-H., 1997. Generation of 
3D city models from Digital Surface Models and 2D GIS, in : 
Baltsavias, E.P. , Eckstein, W., Gülch, E., Hahn, M., Stallmann, 
D., Tempfli, K. , and Welch, R. (eds.), 1997. 3D Reconstruction 
and Modelling of Topographic Objects, International Archives of 
Photogrammetry and Remote Sensing, Vol. 32 Part 3-4W2, Inter- 
national Society of Photogrammetry and Remote Sensing, Syd- 
ney, pp. 68 — 76. 
Jibrini, H., Paparoditis, N., Pierrot-Deseilligny, M., and H. 
Maitre, H., 2000. Automatic building reconstruction from very 
high resolution digital stereopairs using cadastral ground plans, 
Proc. 19th International ISPRS Congress, Amsterdam, 2000. 
Jordan, M., Cord, M., and Belli, T., 2002. Building detection 
from high resolution Digital Elevation Models in urban areas, 
in : Kalliany, R., and Leberl, F., 2002. ISPRS Commission III 
Symposium “Photogrammetric Computer Vision”, Graz, Austria, 
IAPRS, XXXIV-3B, pp. 96-99. 
Moons, T., Frere, D., Vandekerckhove, J., and Van Gool, L. 
1998. Automatic modelling and 3D reconstruction of urban 
house roofs from high resolution aerial imagery, in : Burkhardt, 
H. and Neumann, B., 1998. Computer Vision — ECCV'98, 
LNCS 1406, Springer-Verlag, Berlin / Heidelberg / New York / 
Tokyo, pp. 1.140 — 1.425. 
Paparoditis, N., Maillet, G., Taillandier, F., Jibrini, H., Jung, F., 
Guiges, L., and Boldo, D., 2001. Multi-image 3D feature and 
DSM extraction for change detection and building reconstruction, 
in (Baltsavias et al., 2001), pp. 217-230. 
Roux, M., and Maitre, H., 1997. Three-dimensional description 
of dense urban areas using maps and aerial images, in (Grün ef 
al., 1997), pp. 311—322. 
Van de Woestyne, I., Jordan, M., Moons, T., and Cord, M., 2004. 
DEM segmentation and DTM estimation in complex urban ar- 
eas with ReconLab, Tournesol project T2003.05, Second annual 
report, 2004. 
Vosselman, G., and Suveg, I., 2001. Map based building recon- 
struction from laser data and images, in (Baltsavias et al., 2001), 
pp. 231-239. 
ACKNOWLEDGEMENTS 
The authors gratefully acknowledge the support by the bilateral 
Belgian — French Tournesol exchange research grant, project no. 
T2003.05 “Detection, reconstruction and visualization of 3D sur- 
face models from 3D point and line data”, provided by the Min- 
istery of the Flemish Community (Belgium) and the Ministery of 
Foreign Affairs (France). Special thanks and appreciation is also 
expressed to the Institut Géographique National (Saint-Mandé, 
France) for providing digital imagery and laser scan data used in 
the experiments.