Research of object tracking based on soft feature

To solve the problem that traditional object tracking method had disadvantage of too much time consumed by data process and object lost under complex background, a novel approach to object tracking based on the space spectrum and edge spectrum is proposed and verified experimentally. Analyzing precursor features and study the unity and complementary of space spectrum and edge spectrum, research the extraction technology of moving object based on space spectrum and edge spectrum in order to track and forecast information of the moving object in real time. Experiments show that the proposed approach overcomes the problems of data dependence and object occlusion and achieves good tracking results.