Research On Vehicle Detection And Tracking Algorithm Based On Traffic Monitoring

With the number of motor vehicles increased year by year,road congestion,traffic congestion and the traffic accident rat are getting higher and higher.City security has received widespread attention.The traditional monitoring system needs the staff to deal with massive video monitoring data with serious unreliability and hysteresis.Therefore,intelligent video surveillance is paid more and more attention by scientific research scholars and experts,and the detection and tracking of moving objects are the most important two parts.In this paper,the vehicle on the road is taken as the research object,combined with the characteristics of the vehicle moving quickly,a new improved ViBe algorithm based on the background filling algorithm and the frame difference method is proposed.And realizethe multi-moving target tracking based on this.The main work of the paper is as following:First,the research of the target detection algorithm.The commonly used modeling methods,such as the difference algorithm,the mixed Gaussian background model and the vibe detection algorithm,Through the simulation and comparison,the advantages and disadvantages of each detection algorithm are analyzed.In this paper,we choose the vibe detection algorithm with low computational complexity and good real-time,and use this method to realize the detection of the vehicle.Second,proposed a ViBe detection algorithm based on the background filling algorithm and the interframe difference method.In this paper,a kind of background filling method based on image block is introduced to effectively eliminate the influence of moving objects in the background initialization,and realize the effective suppression of ghosting.At the same time,we can use the background embedding method Inter-frame difference method to locate the moving target to improve the efficiency of the background filling algorithm.Finally,Implementation of vehicle tracking based on particle filter algorithm.By studying the tracking method of moving target and combining with the specific application scene,the particle filter algorithm based on RGB color feature is finally selected.The nonlinear non-Gaussian feature of the particle filter algorithm has a significant advantage in achieving multi-objective tracking.By simplifying the particle filter algorithm based on RGB color feature,this paper reduces the computational complexity and makes it possible to realizethe fast movement of the track.