An Automatic Traffic Surveillance System for Vehicle Tracking and Classification

Shih Hao Yu, Jun-Wei Hsieh*, Yung Sheng Chen, Wen Fong Hu

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

5 Scopus citations

Abstract

This paper presents an automatic traffic surveillance system to estimate important traffic parameters from video sequences using only one camera. Different from traditional methods which classify vehicles into only cars and non-cars, the proposed method has a good capability to categorize cars into more specific classes with a new "linearity" feature. In addition, in order to reduce occlusions of vehicles, an automatic scheme of detecting lane dividing lines is proposed. With the found lane dividing lines, not only occlusions of vehicles can be reduced but also a normalization scheme can be developed for tackling the problems of feature size variations. Once all vehicle features are extracted, an optimal classifier is then designed to robustly categorize vehicles into different classes even though shadows, occlusions, and other noise exist. The designed classifier can collect different evidences from the database and the verified vehicle itself to make better decisions and thus much enhance the robustness and accuracy of classification. Experimental results show that the proposed method is much robust and powerful than other traditional methods.

Fingerprint Dive into the research topics of 'An Automatic Traffic Surveillance System for Vehicle Tracking and Classification'. Together they form a unique fingerprint.

Cite this