The Moving Detection And Object Tracking Technology In Intelligent Video Surveillance System

Intelligent video surveillance system is a product which have great improvement on the traditional surveillance system, it use the computer vision technology to analysis the images captured by the surveillance cameras, automatic detect and recognize people and vehicles in scene, and continuously track them, discover the abnormal behavior and alarm in time. Moving detection and object tracking are the key modules in intelligent video surveillance system, and the research about these two modules is the main job of this paper.In the part of moving detection, we introduced the based principle and algorithm steps of several common moving detection technologies, which include the frame difference method, the optical flow method, and the background difference method. Among them, we paid more space to make detail introduce for background difference method, it contained weighted update method, the gaussian method, the gaussian mixture method, and the improved gaussian mixture method. Finally, we did some experiment for every methods, compared their results to prove the superiority of the improved gaussian mixture method.In the part of object tracking, we made detail introduce for Camshift algorithm which is extensive used by people, and we introduced the based theory of Kalman filter. And then, we discussed how to combine the Camshift algorithm and Kalman filter to do object tracking. In addition, we proposed some points to improve this combined algorithm for static background scene. At the end, we did some experiment to prove this improved object tracking method can really improve the robustness of tracking.