The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part Bl. Beijing 2008 
! 
resume the building edge information and make its outline 
clearer and complete. 
3.3 Analysis of Results 
Because lacking of practical and accurate building layout map, 
we can not analyze the accuracy of segmentation results 
quantitatively but to qualitatively judging by eyes. Comparing 
the extracting results with aerial image data (Fig. 9) we find that 
all the larger buildings in this area have been extracted out 
except for few parts of smaller ones, and no ground points and 
vegetation were mis-judged as buildings. The extracting 
accuracy is very high. 
V * 
t 
a Binary image 
Fig.8 Refined buildings 
4. CONCLUSIONS 
This paper has improved the existing filter-segment method so 
to make it applicable even on the city buildings in undulatory 
regions. It is firstly, by means of a new contour-line-based 
surface prediction filtering method, to easily and quickly 
extract DTM and then normalize the DSM. During segmenting, 
it not only added some geometrical parameters, but also 
changed the segmentation sequence and extracted out more 
accurate building geometry by neighbor iterative approaching 
method. Test results show that although contour-line-based 
surface prediction filtering method is not absolutely accurate, 
yet it is very easy to do and can satisfy the majority 
requirements of extracting building in the undulatory region. 
Changing of segmentation sequence did make the segmentation 
parameter more effective. However, this method still has some 
Fig.9 Aerial image 
deficiency like it can not extract out some very small building 
information. Further improving and updating is still necessary. 
REFERENCES 
Weidner U. An Approach To Building Extraction 
From Digital Surface Models(A). in Intematioanal 
Archives of Photogrammetry and Remote Sensing(C). 
Vienna. 1996. 31(B3): 924-929. 
Mayer H, Automatic object extraction from aerial 
imagery: a survey focusing on buildings[J]. 
Computer Vision and Image Understanding, 1999. 
74(2): 138-149. 
Gamba P ,Houshmand B, Digital Surface Models and 
Building Extraction: A Comparison of IFSAR and 
9. 
10. 
11. 
b Surface image 
LIDAR Data[J], IEEE Transactions on Geoscience 
and Remote Sensing, 2000. 38(4): 1959-1968. 
Maas H-G Two algorithms for extracting building 
models from raw laser altimetry data(A). in ISPRS 
Journal of Photogrammetry & Remote Sensing(C). 
1999. 54(2/3): 153-163. 
Brunn A ,Weidner U. Extracting Buildings From 
Digital Surface Models(A). in Intematioanal Archives 
of Photogrammetry and Remote Sensing(C). 1997. 
32. 
Rottensteiner F, Trinder J, Clode S. Data Acquisition 
for 3D City Models from LIDAR Extracting 
Buildings and Roads(A). in Proceedings of the 25th 
IEEE International Geoscience And Remote Sensing 
Symposium (C). Seoul (South Korea). 2005: 
521-524. 
LIU Jingnan,ZHANG Xiaohong, Classification of 
Laser Scanning Altimetry Data Using Laser 
Intensity[J], Geomatics and Information Science of 
Wuhan University, 2005. 30(3): 189-193. 
Rottensteiner F, et al. Building Dectection Using 
LIDAR Data and Multispectral Images(A). in Digital 
Image Computing: Techniques and Applications (C). 
Macquarie University, Sydney, Australia. 2003(2): 
673-682. 
Rottensteiner F ,Jansa J. Automatic Extraction of 
Buildings From LIDAR Data and Aerial Images(A). 
in Intematioanal Archives of Photogrammetry and 
Remote Sensing(C). Ottawa 2002. 34(4): 569-574. 
Sohn G ,Dowman I, Data fusion of high-resolution 
satellite imagery and LiDAR data for automatic 
building extraction[J]. International Journal of 
Photogrammetry and Remote Sensing, 2007. 
62(l):43-63. 
Kilian, Haala, Englich. Capture and evaluation of 
airborne laser scanner data(A). in International 
287