Computer Video Surveillance Technology

Computer intelligent video surveillance system (CIVSS) is one of important high-tech application fields and front topics in Computer Vision domain. It spans many subjects including computer science, machine vision, image engineering, pattern analysis, artificial intelligence, etc. CIVSS can automatically analyze the sequence of images by the methods of computer vision and video analysis. The system can detect, position, recognize and track objects in a moving environment in real-time. Furthermore, it can also analyze and judge the movement of objects. The aims of CIVSS are to understand the meanings of video stream and to explain the scenes comprehensibly, hence to guide actions and make decisions.As to computer intelligent video surveillance, there are still many problems either in theory research or in applications although broad applications in various fields have shown themselves by now. Large numbers of researchers devoted themselves in the area and have already achieved many progresses. Based on existing conclusions, this dissertation describes researches in the key technologies of CIVSS. The main content includes the detection, withdraws, and tracking of moving objectives and so on.Firstly, this article has conducted the research to moving objectives detection technology, which mainly includes the optical flow method, the instantaneous difference method and the background modeling method. Then the classification to the image division method and the study of the traditional image division methods were carried out. In addition to that, the image division measurement was put up. On the base of the current objectives detection and image division method, a new algorithm of objectives detection and image division, which are available in multi-objectives detecting under complex background, was proposed. This article overviewed the objective tracking methods as well, and pointed out that the target predictions is the essential technology to fast and stable objective tracking. Several kinds of target predictions technology has been studied to shrink the objective matching zone and enhance the objective tracking speed. Finally, a video sequence moving objective detection and the tracking algorithm was proposed.