Moving Target Detection And Tracking Of Network Monitoring System

Network video surveillance technology is an indispensable technology of the cause of today's social security. Relatively simple video surveillance system has been widely used in the social life in many areas. But along with the development of computer science, digital image processing, pattern analysis etc. monitoring system was developing to the direction of intelligence. The information of moving things is always important in information of video surveillance. As an important supplement of the monitoring system, the moving target detection and tracking technology has been a key technology.This paper begins with moving objects detection of monitoring system and segmentation of video images. Firstly, this paper in-depth studies common methods of moving object detecting, such as frames differential method, background frames differential method, optical flow. The method which combines frames differential method and background frames differential method is presented by comparing advantages and disadvantages of above three methods. This method adopts double frames to detect and improves results of moving object detecting and flexibility of neighboring environment.This paper uses the middleware idea, the own interface function encapsulates the underlying hardware API function, it makes the program platform-independent, improves its portability and scalability; the object-oriented idea and fit testing methods make the program clear, reduce the workload of the follow-up testing; Finally, through the platform, it implements the testing and indication of moving targets under camera.