ABSTRACT
Intelligent transportation system is a hot research topic in the current urban construction, and vehicle counting problem is a basic part of intelligent transportation system. In this paper, some methods including region extraction, binarization processing, edge detection and so on are applied to improve the traditional frame difference method. Some problems, such as hollow phenomenon occurring in midway detecting between the vehicle stops, are solved. Experimental results show that precision and applicable rates of the interframe difference are improved, and real-time road vehicle counting based on video surveillance images can be realized.
