Home > Table of Contents


Proceedings of 2009 International Symposium on Computer Science and Computational Technology (ISCSCT 2009)

Huangshan, China, December 26-28, 2009

Editors: Fei Yu, Guangxue Yue, Jian Shu, Yun Liu

AP Catalog Number: AP-PROC-CS-09CN005

ISBN: 978-952-5726-07-7 (Print), 978-952-5726-08-4 (CD-ROM)

Page(s): 381-384

Automatic Detection of Vehicle Activities Based on Particle Filter Tracking

Han Huang, Zhaoquan Cai, Shixu Shi, Xianheng Ma, and Yifan Zhu

Full text: PDF


Automatic detection of vehicle activity by computer vision is one of the most important fields of intelligent video surveillance. This paper proposes an approach of automatic detection for visual vehicle activities based on the traffic video. According to the results of particle filter tracking, the geometry information of vehicle movement is abstracted to understand the behavior of vehicle. There are three components of the based particle filter including particle prediction, sampling updating and Bhattacharyya parameter setting. There are three proposed detection models of vehicle activities including breaking, changing-lane driving and opposite-direction driving which are crucial vehicle activities of common traffic accidents. Finally, several traffic videos are tested, and the results indicate the proposed method and model are efficient, high-detection-rate and robust.

Index Terms

Computer Vision; Image Processing; Particle Filter; Video Tracking; Vehicle Activities Understanding

Copyright @ 2009 ACADEMY PUBLISHER All rights reserved