manual. Centers of gravity of circumscribed rectangles were set 
as the vehicle positions. Figure 6 illustrates one of the 
comparisons of trajectories. The vehicle movement of this 
example is left-hand turn, and the difference between positions 
by the proposed method and manual is 2.6 pixels for this 
experiment. 
Position accuracy of 40 vehicles for 10 seconds was examined, 
and the average of the differences is 2.5 pixels (+0.9 pixels). 
Note that the sequential images of interest was only aligned not 
calibrated absolutely, so the absolute distance of the differences 
were not calculated. 
5. CONCLUSION 
The conclusions of this paper are as follows: 
(1) We evaluated possibility that existing methods may be 
applied; 
(2) We proposed spatio-temporal clustering method as a 
vehicle manoeuvres recognition technique; 
(3) We confirmed the effectiveness of the proposed method 
through 
(a) application to aerial HDTV images, 
(b) application to various different spatial and temporal 
resolution images, 
(c) evaluation of the vehicle position accuracy. 
The future prospects are 
(a) expansion of application; 
(b) absolute orientation to flood of image sequence; 
(c) improvement of accuracy of vehicle position; 
(d) observation by multiple high altitude platforms with video 
cameras 
(e) data fusion of various sensors such as beacons, GPS, fixed 
and high altitude video cameras, and so on. 
The use of this study has the potential to improve many 
Intelligent Transportation System strategies. 
ACKNOWLEDGEMENTS 
We gratefully acknowledge financial support by the Research 
Fellowships of the Japan Society for the Promotion of Science 
for Young Scientists (06958). 
REFERENCES 
Ben-Ezra, M., Werman, M. and Bar-Yam, Y., 2001. A self 
stabilizing robust region finder applied to color and optical flow 
pictures. Image and Vision Computing, 19(7), pp.427-433. 
Betke, M., Haritaoglu, E. and Davis, L.S., 2000. Real-time 
multiple vehicle detection and tracking from a moving vehicle. 
Machine Vision and Appications, 12, pp.69-83. 
Cressie, N.A.C., 1993. Statistics for Spatial Data. John Wiley 
& Sons, New York, pp.105-210. 
Cucchiara, R., Piccardi, M. and Mello, P., 2000. Image analysis 
and rule-based reasoning for a traffic monitoring system. IEEE 
Transactions on Intelligent Transportation Systems, 1(2), 
pp.119-130. 
Fathy, M. and Siyal, M.Y., 1995. An image detection technique 
based on morphological edge detection and background 
differencing for real-time traffic analysis. Pattern Recognition 
Letters, 16, pp.1321-1330. 
Fuse, T., Shimizu, E. and Tsutsumi, M., 2000. A comparative 
study on gradient-based approaches for optical flow estimation. 
In: International Archives of Photogrammetry and Remote 
Sensing, Amsterdam, The Netherlands, Vol.XXXIII, Part B5, 
pp.269-276. 
Fuse, T. and Shimizu, E., 2000. A new technique for vehicle 
tracking on the assumption of stratospheric platforms. In: 
International Archives of Photogrammetry and Remote Sensing, 
Amsterdam, The Netherlands, Vol.XXXIII, Part B5, pp.277- 
284. 
Haritaoglu, IL, Harwood, D. and Davis, L.S., 2000. W*: Real- 
time surveillance of people and their activities. [EEE 
Transactions on Pattern Analysis and Machine Intelligence, 
22(8), pp.809-830. 
Irani, M. and Anandan, P., 1998. A unified approach to moving 
object detection in 2d and 3d scenes, IEEE Transactions on 
Pattern Analysis and Machine Intelligence, 20(6), pp.577-589. 
Kamijo, S., Matsushita, Y., Ikeuchi, K. and Sakauchi, M., 2000. 
Traffic monitoring and accident detection at intersections. 
IEEE Transactions on Intelligent Transportation Systems, 1(2), 
pp.108-118. 
Medioni, G., Cohen, I., Bremond, F., Hongeng, S. and Nevatia, 
R., 2001. Event detection and analysis from video stream. 
IEEE Transactions on Pattern Analysis and Machine 
Intelligence, 23(8), pp.873-889. 
Sangwine, S.J. and Horne, R.E.N., 1998. The Color Image 
Processing Handbook. Chapman & Hall, London. 
Smith, S.M. and Brady, J.M., 1994. A scene segmenter; visual 
tracking of moving vehicles. Engineering Applications of 
Artificial Intelligence, 7(2), pp.191-204. 
Smith, S.M. and Brady, J.M., 1997. SUSAN-A new approach 
to low level image processing. International Journal of 
Computer Vision, 23(1), pp.45-78. 
—566— 
  
KE) 
ABS 
It is 
Syste 
taker 
mod 
its h 
textu 
The 
such 
Unde 
CG1 
appli 
cons! 
have 
two 
imag 
meth 
reali: 
neces 
The 
build 
This 
mode 
scene 
of m 
textu 
came 
Exter 
impr 
the te 
imagi 
city 1 
digit: 
of a 
unifo 
Wher 
becor 
sky u 
aerial 
build: 
from 
using 
upper 
can b 
gaine 
paper