Accurate Vehicle Tracking by Shadow Removal in the Transformed Space

Chang Liu, George Chen, Yingdong Ma, Yajun Ding, and Jinwu Wu

Keywords

perspective transform, background subtraction , shadow removal , vehicle tracking, computer vision

Abstract

In Intelligent Transportation System (ITS), vehicle tracking is necessary to permit high-level analysis, such as vehicle counting or classification. This paper addresses automatic detection and tracking of vehicles in monocular image sequences of traffic scenes recorded by a stationary camera. In order to improve the tracking precision, in the vehicle detection phase, perspective transform and shadow removal are applied to reduce occlusion, which is one of the main reasons for tracking errors. In the vehicle tracking phase, mean-shift with foreground mask is used for building the correspondence between vehicles detected at different time instants. Experimental results from traffic scenes demonstrate the effectiveness and robustness of these methods.

Important Links:



Go Back