The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B3b. Beijing 2008 
533 
• ■ 
■ 
■ *■ 
.« » 
a 
1 
1 ** 
1 
« 
« 
Figure 8. Filtering Binary Vehicle Imagery 
The edges of the extracted vehicles are delineated using the 
thinning algorithm discussed above, and the results are shown 
in Figure 9. 
Figure 9. Outlines of Extracted Vehicle 
In order to illustrate the accuracy, the outlines of the extracted 
vehicles are overlaid on the original imagery, see Figure 10. In 
the overlay imagery the thin red lines indicate the vehicle 
outlines. It can be obverted from Figure 10 that most vehicle 
outlines match well the vehicles. 
ACKNOWLEDGEMENTS 
The research was partially supported by a Natural Sciences and 
Engineering Research Council of Canada (NSERC) discovery 
grant. The PhD research scholarships provided by the Public 
Safety And Emergency Preparedness Canada Research 
Fellowship, the Ontario Graduate Scholarship and School of 
Graduate Studies at University of Waterloo are also 
acknowledged. 
REFERENCES 
Angel, A., M. Hickman, P. Michandani and D. Chandnani, 
2003. Methods of analyzing traffic imagery collected from 
aerial platforms, IEEE Transactions on Intelligent 
Transportation Systems, 4(2): 99-107. 
George, J. K. and Y. Bo, 1995. Fuzzy Sets and Fuzzy Logic: 
Theory and Applications, Prentice Hall PTR, Upper Saddle 
River, New Jersey, USA. 
Goldberg, D. E., 1989. Genetic Algorithms in Search, 
Optimization a.nd Machine Learning, Addison-Wesley. 
Grejner-Brzezinska, D., C. Toth and E. Paska, 2007. Airborne 
remote sensing supporting traffic flow estimates, C. V. Tao and 
J. Li (eds) Advances in Mobile Mapping Technology, Taylor & 
Francis, London. 
Kumar, R., R., H. Sawhney, S. Samarasekera, S. Hsu, H. Tao, 
Y. Gao, K. Hanna, A. Pope, R. Wildes, D. Hirvonen, M. 
Hansen, and P. Burt, 2001. Aerial video surveillance and 
exploitation,” Process IEEE, 99: 1518-1539. 
Pratt, W. K., 1991. Digital Image Processing, Wiley, New York, 
USA. 
Figure 10.Overlaying Vehicle Outlines on Original Imagery 
5. CONCLUSIONS 
In this paper, colour related information is used for imagery 
segmentation purposes. The proposed algorithm combines 
fuzzy c-partition and the genetic algorithm to find fuzzy c- 
partition matrix. 
In order extract the vehicle from UAV aerial imagery, we use 
the uniform radiation of the roads and geometric feature of 
vehicles. 
Several experimental examples show the applicability of this 
approach to vehicle information extraction and the information 
can be used for traffic flow computation and vehicle 
classification.