The Research And Design Of Target Detecting And Tracking System Based On DSP

The intelligent video surveillance has played a very important role in the security of many occasions, such as stations, docks, schools, financial institutions and military area. It is a key factor for the intelligent video surveillance system to finish detecting and tracking the target. This part is closely related to the running performance of the system. Today’s intelligent video surveillance system is mainly built in the computer. Its cost is high, the real-time performance is poor and miniaturization is hard to realize. While the intelligent video surveillance system based on the embedded devices with the DSP processor shows much better results. It has a lot of advantages, such as low cost, high performance and high flexibility. Therefore, this paper is about my work to design and implement the system of detecting and tracking the target based on DSP. The specific content is as follows:First, this paper introduces the development of detecting and tracking the target at home and abroad, including the theory and the engineering application results by the related organizations and high schools. The main problems on detecting and tracking the target are also presented briefly.Secondly, this paper uses the Harris detection operator to extract feature points and uses the matching search criteria based on the three-step search method to complete the feature points matching. Then the Random Sample Consensus method is introduced to estimate the affine motion model parameters. Motion estimation and compensation algorithm simulation test are finished in Matlab.Again, through analyzing and comparing three motion detection algorithms: optical flow method, background difference method and frame difference method, the frame difference method is adopted to detect the moving targets in the image sequence. The binarization and morphological filtering are used for processing the difference image. This paper completes the segmentation of moving targets from the background through the above operations. The target detecting algorithm simulation test is finished in Matlab.Then, this paper analyzes the algorithm of the classical targets match based on feature template, and uses the two-times calculating correlation coefficient method to speed up the targets match. The Kalman filter is used for estimating the target motion’s trajectory to promote the stability of track. The target match and motion’s trajectory algorithm simulation test are finished in Matlab.Finally, this paper builds the target detecting and tracking system based on DSP development platform, and complete the algorithm’ transplantation of the target detecting and tracking from Matlab to DSP. The system is operated to verify the feasibility of the whole software programs. At the end of the paper the system results are presented and analysed.