A robust combined algorithim of object tracking based on moving object detection

In this paper we present a novel combined tracking algorithm based on moving object detection technology. Particle filtering can work well when the object gets an occlusion, it has difficulty in satisfying the requirement of real-time computing. Meanshift can solve this problem easily, it has poor rubustness during mutual occlusion. In the meantime, because backgrounds of many scenes include complex moving objects, many tracking methods only using the color, texture and shape feature of the object often have a poor rubustness. Aiming at all above problems, we detect the moving object using codebook based background model and incorporate the robustess of particle filter with the efficiency of the meanshift algorithm. The experimental results show that our method can handle scenes containing moving backgrounds or illumination variations, and it improves the robustness when the target is occluded.