Multiple Target Tracking With Video Segmentation And Particle Filter

Techniques of tracking multi-target are widely applied in ballistic missile defense, airborne early warning, airborne attack among targets, battlefield surveillance, safeguard of city, human-computer interaction in virtual-reality. Following the development of technology and robotics, technology of tracking multi-target based computer vision becomes a hot research topic in top institutions of various countries.Particle filter is an effective way of object tracking; however, it is still error-prone when it is used for multi-target tracking, especially when multiple objects overlapped. To overcome this problem, we propose to combine particle filter and video segmentation, the latter can provide us with information to distinguish different objects, which is of great help for correct tracking.In chapter one, we introduce the fundamental concept as well as the related research works in multi-target. Our contribution in the paper is also summarized.In chapter two, the application of particle filter in tracking target is investigated.In chapter three, we adopt the technique of video segmentation and particle filter to develop a new algorithm of multi-target tracking.In chapter four, we show some experimental results of the proposed algorithm.Chapter five gives my conclusion and highlights some future work. Keywords: Target Model Video Segmentation Particle Filter Multi-target Tracking...