:>me false edges 
ner type can be 
t. Obviously, 
ie matrix. To 
a search is 
1 for the next 
by scanning 4- 
le desired node 
down comer) 
iowever, in the 
is likely that 
the required 
al tag sequence 
quences and the 
with largest 
ce. 
result 
4 
ed out in the 
1 imagery. In 
analysis and 
ledian filter is 
d point which 
e same as the 
. If the pixel 
*e, the region 
ry precise. In 
ery important, 
ay be not the 
*ure.7. 
(b) 
Figure.7 (a) threshold is 8 (b) threshold is 5 
(b) 
Figure.8 (a) (b) Extraction results 
5.2 Conclusion 
In this paper, we suppose a robust and semi-automatic approach 
to deal with building extraction. This approach is not restricted 
by the shape of the building. It can precisely extract the 
boundary of rectangular building with homogeneous flat 
rooftop with a very low computation complexity which is 
important in data production. The second schema (matrix search) 
supposed introduce a new method to represent shape of 
rectangle or combination of rectangle. Pivotal reason is that it 
depends on the robustness of Hough transform and utilizes a 
new mathematic model to represent shape. However, it has 
some disadvantages; for example, it heavily depends on the 
approximate shape derived via grown algorithm. If the grown 
result is not satisfactory, the extraction may be failed in a large 
extent. Several aspects of the proposed scheme need further 
research. Adaptive region grow approach needs to be explored 
to improve the detection of the building approximate shape. 
Some other characteristics and processes should be incorporated 
into this scheme such as image segmentation. In the future 
research, these issues should be addressed. 
ACKNOWLEDGEMENT 
This work is carried out under the Project for Young Scientist 
Fund sponsored by the National Natural Science Foundations of 
China (40401037) and National Key Basic Research and 
Development Program of China (2006CB701303).The author 
would like to thank Dr Xiangguo, L., and Na, J., for their 
inspiration and encouragement. 
REFERENCE 
Grün, A., Kuebler, O., and Agouris, P., 1995. Automatic 
Extraction of Man-made Objects from Aerial Space Image (I), 
Birkhaeuser Verlag, Berlin, Germany, pp. 321. 
Grün, A., Baltsavias, E., and O, Henricsson., 1997. Automatic 
Extraction of Man-made Objects from Aerial Space Image (II), 
Birkhaeuser Verlag, Berlin, Germany, pp. 393. 
Baltsavias, E., Grün, A., and L, V, Gool., 2001b. Automatic 
Extraction of Man-made Objects from Aerial Space Image (III), 
A. A. Balkema Publishers, Lisse, The Netherlands, pp. 415 
Mayer, H., 1999. Automatic object extraction from aerial 
imagery a survey focusing on buildings, Computer Vision 
and Image Understanding, 74(2), pp. 138 - 149. 
Mohan, R., Nevada. R., 1989. Using perceptual organization to 
extract 3-D structures, IEEE Trans, Pattern Analysis and 
Machine Intelligence, 11, pp. 1121-1139. 
Huertas, A., Nevada, R., 1988. Detecting buildings in aerial 
images, Computer Vision, Graphics, and Image Process, 41, 
pp.131-152. 
Huertas, A., Mohan, R., Nevada, R., 1986. Detecting of 
complex buildings in simple scenes, Inst. Robotics and 
Intelligent Systems, University of South California, Tech. Rep. 
IRIS 203. 
Lee, D. S., Shan, J., and James. S. B., 2003. Class-guide 
building extraction from IKONOS imagery, Photogrammetric 
Engineering& Remote Sensing, 69(2), pp. 143-150. 
Sohn, G., Dowman, I. J., 2001. Extraction of building from 
high resolution satellite data, Automatic Extraction of Man 
made Objects from Aerial Space Image (III), A. A. Balkema 
Publishers, Lisse, The Netherlands. Pp.345-355.