The Research On Vessels Target Detection And Handling Blocking Base On CCTV

Now, CCTV system is playing an important role in the VTS system, such as in the protection of navigational environment safety and unimpede,to improve the efficiency of the motion of the vessels andreduce the probability of maritime accidents,to improve the efficiency of maritime regualtory amd so on. However,in practical applications,there are many related issues still need to solved. For example,when the blocking phenomenon hanppens in the process of tracking the vessels,such bolcking phenomenon would effect the vessl’s locating and tracking. Therefore,we need to study vessels detection and tracking technology especially tracking vessels under the condition of blocking,to ensure the CCTV system could become to be an effective aids tool of the VTS system.In this paper,the moving vessels detection methods are studied based on CCTV system. By compare with frame difference method and background difference method,and comprehensive their respective advantanges,the paper presents a moving vessels dectection method combine three-frame difference method with backgroud difference method. The experimental results show that this method can detect vessel targets effective,it provide a target for the subsequent vessel target tarcking.Then,we study vessel target tracking techology including MeanShift algorithm and CamShift algorithm,based on the vessel targer was detectived. By analyzing the CamShift algorithm applied in vessel targer tracking,we found that this algorithm would loss the target when the background is under the same color interference. Therefore,this paper introduces the Kalman filter. The Kamlan filter is used to predict the position of the vessel target in order to solve the problem during track the vessel target. According to the blocking phenomenon in the process of tracking,this paper discuss the blocking phenomenon judgment strategy and we judgment the blocking phenomenon by using convex hull information of the vessel target. After the blocking phenomenon was detectived,the SIFT matching algorithm was studied in this paper. This algorithm is an effective object macthing method which can be realized macth the vessel target under the blocking phenomenon. By these two kinds of algorithms to deal with the blocking phenomenon during track the vessel target. The experiment proved that this algorithm has obvious improvement compared with the traditional tracking algorithm. It is still able to achieve accurate tracking the vessel target under the same color background or when the blocking phenomenon was happened.