The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B5. Beijing 2008 
650 
ranks above. During the experiment it has separately done the 
simplification in three ways under the single condition and got 
three different simplified models of their own effects. These 
three simplifying strategies have some diversity in both 
simplifying effect (Figure 5) and efficiency (Table 2), which 
give an opportunity for realizing personality to the clients and 
own a good flexibility. 
(a) (b) (c) 
Figure 5. The effect of 3 simplification strategies 
\name 
stair\ 
Latency 
triangle 
number 
simplified 
triangle 
number 
Simplified 
rate 
(%) 
remark 
First 
stair 
97,266 
11,074 
11.39 
only for the 
triangles of 
heavy 
Second 
stair 
97,266 
27,522 
28.30 
for heavy 
and 
moderate 
Third 
stair 
97,266 
46,296 
47.60 
for ail 
Table 2. The comparison for different strategies’ efficiency 
The experiment indicates that: 
• In the aspect of simplification effect, the model 
simplification driven by perception can well keep the 
surface characters of the components, which ensures the 
consistency on the surface of the whole building before and 
after the simplification. 
• In the aspect of efficiency, together with the effort of 
simplification threshold value, the simplification operation 
driven by perception has sharply cut down the whole model 
data amount, and the model simplified rate is ranging from 
10% to 50% according to different simplification strategies. 
• For the LOD model of the highest level which needs to be 
built, it can sharply reduce the model’s data volume; 
meanwhile well keep the shape of the whole model. 
• Those three kinds of simplifying strategies in the process 
offer many choices to the clients, and they can flexibly 
choose a suitable way according to their hardware and soft 
ware conditions and their grasps to the granularities for 
model representing. 
6. CONCLUSION 
In currently, complex building models are widely used in 
various fields, such as VGEs, cyber city, and culture heritage 
documentation, with well development of the virtual reality and 
3DGIS technologies. But such high complexity and huge data 
volume of building models can not meet the real-time and 
interactive application requirements in virtual environments. 
Perception-driven simplification in this paper is founded on the 
image process that rendered image of building models are 
filtered by HVS filter based on wavelet transformation. The 
perceptual information is acquired and passed to model surface 
by ray casting through finding the difference between rendered 
image and filtered image. The model surface’s perceptual 
information drives simplification operation directly to reduce 
the imperceptible details on model surface. 
A prototype is developed to create multiple levels of building 
models. The simplification experiment of a complex ancient 
building model proves that the simplified models not only 
preserve the perceptually details of the model, but also 
minimize the amount of triangles and keep high geometric 
fidelity. 
REFERENCES 
Butler, D., 2006. Virtual Globes: The Web-wide World. Nature, 
439(2), pp. 776-778. 
Coors, V., 2001. Feature-preserving Simplification in Web- 
based 3D-GIS. In: Proceedings of First International 
Symposium on Smart Graphics, New York, USA. 
Forberg, A., 2004. Generalization of 3D Building Data Based 
on a Scale-Spaces Approach. In: International Archives of the 
Photogrammetry, Remote Sensing and Spatial Inforamtion 
Science, Vol.XXXV, pp. 194-199. 
Forberg, A., 2007. Generalization of 3D Building Data Based 
on a Scale-Spaces Approach. ISPRS Journal of 
Photogrammetry and Remote Sensing, 62(2), pp. 104-111. 
Forberg, A. and Mayer, H., 2002. Generalization of 3D 
Building Data Based on a Scale-Spaces Approach. In: 
Proceedings of the ISPRS Technical Commission IV Symposium 
2002 on Geospatial Theory, Processing and Applications, 
Ottawa, Canada, Vol.IAPRS, XXXIV, part 4, 6p. 
Kada, M., 2002. Automatic Generalisation of 3D Building 
Models. In: Proceedings of the ISPRS Technical Commission IV 
Symposium 2002 on Geospatial Theory, Processing and 
Applications, Ottawa, Canada, Vol.34, 6p. 
Kada, M., 2007. A Contribution to 3D Generalization. In: 51st 
Photogrammetric Week, Stuttgart, Germany, pp. 41-51. 
Kobbelt, L., Campagna, S. and Seidel, H., 1998. A General 
Framework for Mesh Decimation. In: Proceedings of the 24th 
Conference on Graphics Interface (GI-98), Vancouver, Canada, 
pp. 43-50. 
Mannos, J.L. and Sakrison, D.J., 1974. The Effects of a Visual 
Fidelity Criterion on the Encoding of Images. IEEE 
Transactions on Information Theory, 20(4), pp. 525-535. 
Meng, F. and Zha, H., 2004. An Easy Viewer for Out-of-Core