Multi-moving Targets Tracking Based On The Video

One of the Computer Vision task is moving object detection and tracking, it combines with the image processing,pattern recognition,automatic control,computer and many other areas of artificial intelligence. It is mainly used in military visual guidance,video monitoring,medical diagnostics,intelligent transportation and other aspects. Target tracking technology has been used in many important areas, so the technology has become a major research.Moving object detection and extraction based on the static background and moving target tracking aspects are studied in the paper.(1) Firstly, the commonly used motion detection algorithm is introduced, such as frame difference algorithm and background subtraction. The Gaussian Mixture Model algorithm is used to extract the background, which is described. In the experiment, the Gaussian Mixture Model algorithm and edge detection algorithm are adopted.(2) On the aspect of the moving target tracking, the basic knowledge of the mean-shift algorithm is introduced. Foreground detection and mean-shift algorithm are both adopted. Looking for the real position by calculating the mean-shift vector. Experiments verify that the accuracy of vehicle tracking is improved by using the algorithm(3) Draw the trajectory tracking of vehicles, when the vehicle occurring, you can discover illegal vehicles.Because of the needs of the laboratory project, the code of the tracking has been extracted, which is convenient for later using.