A method of Multi-targets detection and tracking from traffic video based on particle filtering

摘要

In this paper, a framework for detection and Multi-targets tracking problems by using particle filter is developed. The characteristics of traffic video, structural of road profile and movement of vehicles have been analyzed. In order to overcome the disadvantages of artificial initialization in most of the target tracking algorithms, we designed a method of detection based on the characteristics of road and vehicle. Aims to reduce the operation complexity, as well as improve the accuracy and rapidity, the heuristic method for multi-targets data correlation based on the characteristics of video sequence is proposed. In the end, we integrated rapid detection with data correlation method, and designed the multi-targets tracking algorithm based on particle filter for traffic video sequence. Experimental results demonstrate that target can be fast detected in the required time by the detection method. The method of tracking based on particle filter has good tracking performance in traffic video and is effective in tracking multi-targets.