Research And Implementation Of Real-time Tracking Technology For Moving Object Under Complex Scene

With the constant advancement of video processing technology,video surveillance has the intelligent characters. As a key technology of intelligent video surveillance, the moving object tracking technology should be studied. Most researches only focus on the simple and static background scene, but rarely on complex scene.Along with the wide-range application of the intelligent video surveillance in complex scene,there are two problems should be taken in, which are how to gain the good tracking quality and how to implement efficiently the object tracking system. This paper designed the tracking system architecture which adapt in complex scene. In this system, the embedded terminal is use to detect moving object, and moving object template matching is implemented in the video terminal. The experiment results demonstrate the effectiveness of this architecture.We proposed a moving object tracking approach which been in distributed collaborative style. By the approach, the computing resources could be utilizated effectively, and the accuracy and effectiveness of tracking could be improved.For the complex scene , we research the approach based on wavelet transform and template matching for tracking moving object. The approach is consisted of the object detection algorithm based on embedded system, and the object tracking algorithm based on PC computer environment. It can enhance the speed and accuracy of object tracking in complex scene.In this paper, an algorithm of camera movement control is presented based on the pan-silt closed-loop control system structure. It can resolve the hysteretic nature of object tracking system.Experiment results show that the moving object tracking approach has the virtues of fast computing speed, it can meet the real-time requirements, and is robustious.