A Moving Object Detection Algorithm for Vehicle Localization

This paper proposed a video-based moving object detection algorithm for vehicle localization. First, construct a background image, and then extraction moving object through the difference between current image and background image. This paper also combined concept of region of interesting to eliminate noise interference in other regions. the experimental results show that the proposed algorithm is efficacious and reliable in vehicle localization. since the vehicle is localized, it can be not only calculated number of vehicles, but also computed average vehicle speed according to difference of time and distance. This real-time traffic information could be consulted for traffic control.