Space Moving Target Trajectory Tracking System

Along with the development of technology, computer vision has become a frontier science of the most popular topics, and also have more access to people's lives. Intelligent video surveillance is an important computer vision components, the core is to handle the real-time scene of all kinds to get the information of the moving targets, such like shape, location, color and so on with no human intervention in the case. Then do the analysis, judgment, and forecast on the moving object on the base of the method mentioned above to achieve the people's expect of purpose.On the current domestic there are many researches on the tracking of the moving objects and also have achieved good results. But it is very difficult to find a broad applicability of the method, and the prices of the system are comparatively high, more over, the algorithm is not such optimization.This paper proposes a hybrid model based on adaptive multi-target tracking and detection algorithm. On the process of the noise, the method of corrosion and expansion obtain more precise moving targets. When tracking the target, introduction of the moving target measure of reliability enhance the tracking of stability and reliability on the base of the target block algorithm of the various possibilities of analysis. Second, when moving targets surveillance is beyond the current scene, serial communication control achieves the tracking of the scene which contain the moving objects, third, this paper founded a real-time tracking system and did many experiments.In the actual scene of a number of video sequences of the experiment, the system shows a strong adaptability and good real-time processing. It can handle the various situations encountered in dairy life and have a good performance.