The research of video browsing technology based on the moving target

This study is designed to achieve a video browsing technology which allows users to browse the moving target of interest by controlling the moving object in the video scene. This study included the following: first, the tracking accuracy needed to be improved; second, we needed to accelerate the tracking speed to ensure that users could view real-time changes of moving object; third, a user-friendly interactive interface was needed. In this study, we applied both the pyramid Lucas-Kanade optical flow algorithm and Camshift algorithm to track the moving target, and optimized both of them by using parallel processing algorithm. Experimental results show that Camshift algorithm can make up the insufficient of optical flow algorithm to a certain extent, and the optimized algorithms are able to achieve real-time browsing of high-definition video.