of Plane 
  
convergence 
The process 
xecuted until 
consequently 
js to reflect 
'd area-based 
| preliminary 
(a. The aerial 
200 dpi. The 
parallax (see 
M model and 
mpared and 
teristics. 
image in the 
rapping, shift 
or comparing 
alculated with 
g result. The 
based on its 
or comparison. 
- experiments. 
ts, and Figure 
d 13 illustrate 
ap) of typical 
vith enhanced 
illustration of 
yme buildings 
de and sharp 
| of enhanced 
ig results have 
|. model where 
  
| image 
  
(a) Left image (b) Right image 
Figure 10. Detected and matched edges 
uz M Va = ta ex 
(a) Shift vectors (b) Depth map 
Figure 12. Processing results of typical ANM 
    
(a) Shift vectors (b) Depth map 
    
(a) Case of typical ANM 
Figure 14. Enlarged illustrations of depth map 
   
  
    
  
    
    
     
Figure 13. Processing results of enhanced multi-cluster ANM 
(b) Case of enhanced 
multi-cluster ANM 
6. CONCLUSIONS 
In this paper, we have proposed a stereo matching technique 
with improved Adaptive Nonlinear Mapping by area-based 
multi-cluster model. Preliminary experimental results showed 
the effectiveness of the proposed approach for reconstruction of 
DSM where plane approximations were adequately processed. 
In this study only horizontal plane approximations were 
considered for clustering. The proposed approach is also 
capable of further improving stereo matching in urban area by 
considering planes of multiple directions. 
REFERENCES: 
Horn, B. K. P. 1986. Robot Vision, MIT Press, Cambridge, MA. 
Kosugi, Y., Sase, M., Kuwatani, H. et al., 1993. Neural network 
mapping for nonlinear stereotactic normalization of brain MR 
images. Journal of Computer Assisted Tomography, Vol.17, 
No.3, pp.445-460. 
Kosugi, Y., Fukunishi, M., Sakamoto, M., Lu, W. and Doihara, 
T., 2001. Detection of sheer changes in aerial photo images 
using an adaptive nonlinear mapping. Proc. DMGIS 2001, 
pp.145-146, Bangkok. 
Marr, D. and Poggio, T. 1979. A computational theory of 
human stereo vision. Proc. Royal Soc. London., vol. B204, 
pp.301-328. 
Moravec, H.P. 1981. Robot rover visual navigation. UMI 
Research Press, Ann Arbor, MI. 
Sakamoto, M., Uchida, O. and Wang, P., 1998. Automatic high 
accuracy tie points detection in stereo images using relaxation 
method and gradient method. Journal of the Japan Society of 
Photogrammetry and Remote Sensing, Vol.37, No.5, pp.35-46. 
Sakamoto, M., Lu, W., Wang, P. and Kosugi, Y., 2001. A new 
stereo matching approach using edges and nonlinear matching 
process objected for urban area. Proc. DMGIS 2001, pp.248- 
253, Bangkok. 
Smith, S.M. and Brady, J.M., 1997. SUSAN -A new approach 
to low level image processing. Int. Journal of Computer Vision, 
Vol.23, No.1, pp.45-78. 
—89—