Video Vehicle Detection Algorithm based on Virtual-Line Group

The paper presents the video vehicle detection algorithm based on virtual-line group. The algorithm uses virtual-line group to optimize the states of detection lines in both spatial domain and temporal domain by setting the domain-knowledge based rules. By synthesizing the optimization information in the larger spatio-temporal scale, the algorithm can improve the precision of vehicle detection. The experiment results and the detailed analysis demonstrate that the algorithm is realtime, accurate and robust