The Study Of Methods On Traffic Events Detection By Video Images Analysis

The main function of the Traffic Events Detection System (TEDS) is able to detect and deal with the traffic events on roads so as to decrease the death rate and the loss of wealth. Moreover, TEDS can help avoid sequent events, save energy sources, reduce pollution and so on. As the core of the TEDS, algorithm of traffic events detection is worth studying.According to the fundamental requirements of traffic events detection, techniques and theories relevant to traffic events detection are systematically investigated in this thesis. The main contents can be stated as follows:First, this thesis introduces the correlative techniques and research of traffic events detection, and then analyzes the class and evaluation of the algorithm.Segmentations based on frame-difference and accumulative difference images are realized. The edges of the results of background-difference are picked up. Expansion is operated on the results for 3, 5,7,10 times with 3×3 structure elements in morphology. Then some methods of video images analysis are discussed in detail.Thorough discussions are made which are about the detection of directions and speed of the vehicles based on tracking. A new algorithm using texture characteristic is presented which aims at resolving the problems of the tracking algorithm which is complex and time consuming. Some experiments prove that the new algorithm can provide real-time traffic conditions of the road and detect the crowd efficiently. At last, algorithm about collision incidents detection based on Markov random field is discussed.