Study on Moving Object Tracking Algorithm in Video Images

摘要

Moving object tracking algorithm in video images is studied. Mean shift is optimization algorithm based on gradient descended, which tracks moving object through iteration. While particle filter is a robust method of tracking in non-gauss and non-linear case. Considering the excellence and disadvantage of mean shift and particle filter algorithm, a tracking algorithm based upon the two methods is proposed. Experiments show that the method can tracking moving objects efficiently, even in the situation of occlusion.