Video Tracking Algorithm Applied Research In The Mean-shift

Video target tracking technology is an important research topic in the field of computer vision, it combines computer science, pattern recognition, artificial intelligence and some other scientific knowledge and it is widely used in industry and national defense. Mean-Shift algorithm is an excellent algorithm in video target tracking, which is based on nonparametric estimation theory and the target color characteristics model. The algorithm has the advantages of small amount of calculation and robustness, but unfortunately, the Mean-Shift algorithm won't perform well in a complex environment, so it's necessary to make improvement to the Mean-Shift algorithm.The main contributions of the paper are as follows:(1) Kalman Filter is introduced to improve the Mean-Shift algorithm, which is helpful for tracking the target in high speed. And then, Kalman Filter is also improved in the paper.(2) In the tracking process, in order to adapt for the target changes, a partial template update algorithm is applied to make the template accurate.(3) Due to the interference of color target in the background, a method of applying the color information in the template is adopted to reassess the current target pixels, which exclude the interference of background.(4) For the occlusion of the target, a local and overall scheme is used to find the non-obscured target message, which combined with objective position can locate the target, thus making the target is being tracked.(5) Based on the Mean-Shift algorithm, the video detecting approach is introduced in this paper. A method of background subtraction is used to identify the movement targets in the pictures. Then the Mean-Shift algorithm is used to track the targets and to finish the video tracking system, which completes the detection and tracking task. Finally, a software system of target detecting and tracking is developed.