Detecting Moving Targets from Traffic Video Based on the Dynamic Background Model

An efficient method to detect the moving target in traffic video based on the dynamic background model is proposed in this paper, after analyzing existing methods for target detection. The model of target detection is given firstly, then a rough set weighted classification method for video image is presented. Based on the video classifications, the background model is established on the historical data. The background judgment and moving object detection for video are done with this model and then the background model is updated with the current video. The experimental results show that this method can adapt the diversification of background and has high adaptability and precision. The processing speed can meet the requirement of real time detection.