| Vision tracking technology is an important branch of machine vision, which has broad prospects in the field of video surveillance, vision navigation, medical and so on. With the enhancement of sensor technology and computer hardware performance, vision tracking has shown its broad application. This paper has the practical background of border security protecting. The comprehensive border security protecting system based on visual servo has been designed and realized. In addition, tracking algorithm based on detection and particle filter were designed respectively for the system.Firstly, the research status of related technologies is summarized. Considering the practical needs of the subject, the solution of border security system based on visual servo was proposed. The system consists of the host computer subsystem, the intrusion alarming subsystem, the vision tracking subsystem and the flexible tracking subsystem. Design of the vision tracking system was analyzed while the algorithm framework was given which made the foundation for the subsequent algorithm.Secondly, moving target detection algorithms were studied. The detection principle and experiment results of the triple-frames-difference-method and the Lucas-Kanade optical flow method were analyzed, respectively. As there were noise points in the method of frame difference and untrue corners in optical flow method caused by jitter of background, the compensation algorithm based on gray projection was studied and then an experiment was given to demonstrate the effectiveness of the method.In addition, the article also focused on vision tracking algorithm design and implement. Considering the functions of the comprehensive border security system, the single target tracking algorithm and the multi-target tracking algorithm were designed based on the two methods, respectively. The effect and influence factors were analyzed through theoretical analysis and experiments. On multi-target tracking problem, the methods of communication domain detection and joint probabilistic data association were used to distinguish different targets, which can improve its performance effectively.At last, the platform of comprehensive border security system based on visual servo was designed and implemented. The hardware structure of the security system component was given, and visual servo system design and PTZ control algorithm were analyzed. Rotation servo tracking of the visual system was tested, then the influencing factors were analyzed. |
