ee of 
etric; 
Was 
rates 
non- 
hirty- 
Xf the 
lem. 
ours. 
non- 
true 
; and 
plied 
6 as 
ed in 
; the 
with 
  
Zheng Wang 
  
   
Figure 11. Corners, structure and TIN of Building 12. Figure 12. Corners, structure and TIN of Building 15. 
  
  
  
Figure 13. Corners, structure and TIN of Building 14. Figure 14. Corners, structure and TIN of Building 16. 
4. CONCLUSION AND FUTURE DEVELOPMENTS 
The approach has been successfully tested with real world data. The test results demonstrate that the approach is 
capable to extract and reconstruct buildings from high quality terrain surface, such as LIDAR data generated surface. 
The approach utilizes closed edges to analyze object shapes with properties like symmetry and circularity. The 
combination of symmetry, circularity, orthogonality, and parallelism provides very effective and sufficient measures to 
classify buildings. The successful tests with the real LIDAR data on the classification processes demonstrate the 
practical capability of the processes. The TIN-model-based method has shown to be an effective way to generate 
building surfaces with normal, real world data sets, i.e., data sets with some noise. It is the contribution of closed edges 
  
International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B3. Amsterdam 2000. 963