Study On The Key Technology Of Intelligence Surveillance

The key technology, which is used to do intelligence surveillance, has been used widely in all aspects of our life. And the intelligence surveillance, which has a blur definition, is to use the methods of digital image processing and computer's sensing to analyze automatically to obtain the targets'orientation, recognition and to track them. In this process, we use some methods that based on image processing, artificial intelligence, and pattern recognition to get the information we need. It is obvious that the most important thing about this process is the result of identification and the tracking.This paper mainly discusses the fundamental theories and key technologies of Intelligent Video Surveillance, including detection and extraction of moving objects with static background, detection of shadow of moving objects and tracking of moving objects and so on. This paper has spent most of time on the algorithm which is used to obtain the background information. The algorithm is on the assumption that the objects move on a computable speed and the background are always seen. Experiments results given in the thesis show that, the algorithm are simple, fast and robust, which could be used in really intelligence system.The camera motion detection system can be used in many aspects of intelligence surveillance. As it is necessary to track objects on which we take great interest, the camera must move in most of time. For the motion needs to adjust the direction of the cameras, of course, this motion can only be accepted by some types of cameras, it is better that we some manage to get the result of the motion, which means if the camera have moved correctly. What is more, motion information of camera is an important feature which describes video information. Thus, it is very necessary and important to find some way to detection the camera motion.After deeply research on the global motion, this paper presents a real-time system which uses the detection algorithm based on the global motion to detect camera motion. It gets the camera motion information through analyzing the data captured by the camera, and returns the result whether the camera worked well to user. This paper has done many experimental in real environment within many kinds of weather, and a great number of types of traffic conditions. The experimental results show that this system which is designed based on the algorithm could give out the result effectively with high accuracy.