nbul 2004 
  
ES UVP 
ANE 
  
f 
x 
Ln 
3 
wl 
  
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B3. Istanbul 2004 
  
  
(f) 
Figure 6 Image patches before and after shadow removal 
shadow detection. (a), (c) and (e) are some patches of the 
original image. (b), (d) and (f) are the restored image by CAIM 
to (a), (c) and (e) respectively. 
From the experiment images we can see, that most shadow areas 
have been restored successfully. Somewhere the color and 
brightness differ from their surroundings. One reason is the 
chosen of the companion area. It is selected as a neighboring 
region along the shadow direction. If it covers just as same as 
what the shadow area contains, the restoring will be perfect. This 
can be obtained roughly by DSM data. However the companion 
is chosen now by computation on the image without the height 
information, thus companion area may cover some objects not 
existing in the shadow area, and vice versa. But the chosen of 
companion area based on DSM is unpractical for the time being 
because it will bring in much more computation. 
807 
AS conclusion, our research makes several main achievements as 
following. 1. Developed several shadow detection method. a) 
Directly detect and segment the shadow region from the image, 
not using DSM and photogrammetry. b) Compute the cast 
shadow by DSM and the sun altitudes, and trace the shadow to 
the image plane. c) Considering the advantages and 
shortcomings of the two method, proposed an integrate method. 
2. Proposed a set of techniques to remove the shadow region by 
image analysis. a) Companion area to a shadow area is chosen. b) 
Pre-processing of the histograms of the shadow area and its 
companion reduces the impact to the mapping result of the 
lowest intensities and the highest intensities. c) Seamless mosaic 
to smooth the edges of the shadow area after intensity mapping. 
Reference: 
ADS40 Information Kit for Third-Party Developers! | 
www. gis. leica-geosystems.com 
Bittner J., Hierarchical techniques for visibility determination, 
Postgraduate Study Report, DC-PSR-99-05. 
Castleman K.R., 1998. Digital Image Processing, Prentice-Hall 
International, Inc., Press of Tsinghua University, Beijing, China, 
1998. 
Chen, L.C. and Rau, J.Y., 1993. A unified solution for digital 
terrain model and orthoimage generation from SPOT stereopairs, 
IEEE Trans. on Geoscience and Remote Sensing, 31(6), 
pp.1243-1252. 
Paglieroni D. W., 1997. Directional distance transforms and 
height field preprocessing for efficient ray tracing, Graphical 
Models and Image Processing, 59(4), pp.253—264. 
Paglieroni D. W., 1999. A complexity analysis for directional 
parametric height field ray tracing, Graphical Models and Image 
Processing, 61, pp.299-321. 
Paglierroni D. W. and Petersen S. M., 1994. Height distributional 
distance transform methods for height field ray tracing, ACM 
Transactions on Graphics, 13(4), pp.376-399. 
Zhou G., Qin Z., Kauffmann P., Rand J., 2003. Large-scale city 
true orthophoto mapping for urban GIS application, ASPRS 
2003 Annual Conference Proceedings.