International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B5. Istanbul 2004 
  
Figure 8: A complex example of a facade occluded by pedestrians and a statue of Schiller. The top row shows three images taken 
from three different stations. The middle row shows the three images rectified individually. The bottom row shows the three images 
transformed to the plane of the third image. 
5 SUMMARY 
We have demonstrated a simple approach for multi-image fusion 
for the generation of occlusion-free textures for façade texturing. 
Both the cases of moving objects and stationary objects occlud- 
ing a façade can be handled in a unified manner. The approach 
does not use complex 3D computations. Specifically it does not 
require the determination of the exterior orientation of the sensor 
or the use of control points in object space. The approach is suit- 
able for use with uncalibrated cameras since it only requires the 
computation of the perspective projection of images. Arguments 
can be made whether the lens distortions of the camera ought to 
be calibrated. Our experience has shown that the distortions intro- 
duced by the facade's relief during rectification are by far larger 
than the distortions caused by quality lenses. 
Our approach requires the measurement of four points per image 
to be included in the fusion process. Comparing this to the sim- 
ple single image approach for facade texturing, we see that the 
manual work load will increase linearly with the number of im- 
ages. Due to this moderate need of user interaction it holds the 
potential for large-scale use. 
Of course full automation is always desirable. Using a proce- 
dure, which is able to automatically detect buildings in terres- 
trial imagery, will lead to full automation. We have presented 
such a procedure in a previous publication (Bóhm et al., 2002). 
We have described a method for automated appearance-based de- 
tection of buildings in terrestrial images. From an image with a 
given approximated exterior orientation and a three-dimensional 
CAD Model of the building, we were able to detect the exact 
location of the building in the image. The method used the com- 
bination of an image device and some hardware to approximately 
measure orientation. In future work we aim to integrate this pro- 
cedure with the results presented here to achieve full automation 
of multi-image fusion. 
References 
Bohm, J, Haala, N. and Kapusy P, 2002. Automated 
appearance-based building detection in terrestrial images. In: 
ISPRS Commission V Symposium, International Archives 
on Photogrammetry and Remote Sensing, Vol. 34number 5, 
pp. 491—495. 
Bornik, A., Karner, K. E, Bauer, J., Leberl, F. and Mayer, 
H., 2001. High-quality texture reconstruction from multi- 
ple views. Journal of Visualization and Computer Animation 
12(5), pp. 263-276. 
Coorg, S. and Teller, S., 1999. Extracting textured vertical fa- 
cades from controlled close-range imagery. In: IEEE Com- 
puter Society Conference on Computer Vision and Pattern 
Recognition, pp. 625-632. 
El-Hakim, S., Brenner, C. and Roth, G., 1998. An approach 
to creating virtual environments using range and texture. In: 
IAPRS, Vol. 32number 5, Hakodate, Japan, pp. 331-338. 
Fischler, M. A. and Bolles, R. C., 1981. Random sample consen- 
sus: A paradigm for model fitting with applications to image 
analysis and automated cartography. Communications of the 
ACM 24(6), pp. 381-393. 
Haala, N., Bohm, J. and Kada, M., 2002. Processing of 3d build- 
ing models for location aware applications. In: ISPRS Com- 
mission III Symposium, International Archives on Photogram- 
metry and Remote Sensing, Vol. 34number 3, pp. 138-143. 
Kada, M., Roettger, S., Weiss, K., Ertl, T. and Fritsch, D., 2003. 
Real-time visualisation of urban landscapes using open-source 
software. In: Proceedings of ACRS 2003 ISRS, Busan, Korea. 
Prati, A., Mikic, L, Trivedi, M. M. and Cucchiara, R., 2003. De- 
tecting moving shadows: Algorithms and evaluation. IEEE 
PAMI 25(7), pp. 918—923. 
Toyama, K., Krumm, J., Brumitt, B. and Meyers, B., 1999. 
Wallflower: Principles and practice of background mainte- 
nance. In: ICCV99, pp. 255-261. 
Wang, X., Totaro, S., Taillandier, F., Hanson, A. and Teller, S., 
2002. Recovering facade texture and microstructure from real- 
world images. In: Proc. 2nd International Workshop on Tex- 
ture Analysis and Synthesis, pp. 145—149. 
   
  
  
  
   
   
   
  
  
  
   
   
   
  
   
   
  
   
   
   
   
   
  
  
  
    
  
   
  
  
   
   
   
   
  
   
   
   
  
   
  
   
  
   
  
  
  
  
K 
In 
ap 
ph 
of 
sp 
pr 
ad 
pa 
SC: 
fle 
co 
ph 
du 
V] 
po 
mi 
the 
cel 
ref 
of 
tex 
ph 
Th 
cir 
sui 
as 
on 
(th 
pre 
em 
de 
pat 
im;