Research On The Key Technologies Of Compressive Tracking

The target tracking was the computer vision domain important research topic, the target tracking has had the significant influence to humanity’s life, obtained widely has applied.For example the safe monitoring, man-machine interaction, military domain and so on, but the target tracking also has many challenges, for example strength of illumination change, background harassment, mask, questions and so on criterion change.How therefore designs an effective track algorithm to have the necessity extremely, this article main research work is as follows:(1) First we elaborate the target tracking algorithm research background and the significance, as well as the domestic and foreign research present situation, has analyzed the main question and the difficulty which in the target tracking algorithm exists.(2)We introduce two kinds of typical based on the compression sensation theory target tracking algorithm, and design an experiment to confirm its algorithm the validity, The advantages and disadvantages of these two methods were systematically explained.(3) To solve the problem of compresive tracking algorithm track frame window size invariable easy to cause the track drifting in the track process the situation, We analysis quite common window auto-adapted adjustment algorithm, the union compresive tracking algorithm characteristic to propose one kind based on the window auto-adapted adjustment compresive tracking algorithm, this method and has established the track window adjustment mathematical model based on the image criterion information content change, this method can adjust promptly tracks the frame the size, thus enhanced the track algorithm effective line.The experimental result indicated this article proposed the algorithm has obtained the good track effect in the goal criterion change scene.(4) To solve the problem of compresive tracking algorithm to use the sole target characteristic not to be able accurate comprehensive description simulated target this question, through the analysis quite common multi-characteristic fusion method, proposed more than one kind of characteristic auto-adapted fusion compresive tracking algorithm, this method unified the texture and the color characteristic, can in the different scene the auto-adapted adjustment fusion coefficient, thus increased the description simulated target reliability, the experimental result had indicated this article proposed the algorithm has obtained the good effect in many kinds of track scene.