Study On The System For Video Surveillance Based On DAM6416P

In this thesis, the methods of image pre-processing, moving object detecting, segmenting and tracking in video image sequences are studied, referring to advanced image processing technique home and overseas at present. Aim at certain backgrouds, the frame differentiation method is adopted in comparison with other algorithms according to the results of experiments. And a double gate object tracking method based on prearranged points circuiting is presented. Through the corresponding software and hardware designing, a system for video surveillance based on DAM6416P image processing platform is set up.The process of this system is as follows. First, the camera begins to collect image sequences after moving to a certain prearranged point. The image sequences collected are pre-processed through the methods of histogram equalization, median filtering and so on by DSP. Second, the frame differentiation method is adopted to detect moving objects. Once there is on moving object, the camera continues to move to next prearranged point. The other way round, the moving object contour is precisely abstracted by means of thresholding segmentation, morphological filtering and the analysis of image connectivity. Lastly, the double gate centroid tracking method is presented to track the moving object. The movement of the object is fixed on by judging the relationship of the outside gate and the centroid of the object. And then the camera is controlled to move to the corresponding prearranged point so that the moving object is tracked approximatively real-time.The high speed digital signal processor TMS320C6416 is the kernel device of this system, and the software environment is Code Compile Studio. The results of experiments on some specific scene prove the feasibility of the system.