Research On Imaging Detection, Recognition And Tracking Technology Of Moving Object

Moving targets real-time detection, recognition and tracking in complex background is an important research task in Automatic Target Recognition (ATR) field. It has been widely used in military filed such as weapon control and guiding system. According to the current development status of visible light image processing methods and the main problems existed as well as the basic requirements on its practical engineering applications. This thesis studied the key technology of the video tracking system —automatic moving target detection, recognition, tracking and their real-time dealing methods from two angles which are analyzing algorithm and making experiments. The main achievements of this paper are as follows:(1) An adaptive method of automatic threshold selection based on the improved method of Otsu thresholding is given. At the same time, designed a criterion for discrimination if correct segmentation is obtained, and accomplishes the purpose of image segmentation automatically with the use of iteration. The method has a good adaptive, and visible light targets with different dimension can be segmented out of complex background quickly, completely and rapidly;(2) At the aspect of feature matching, the author put forward the feature vector that amalgamated the changeless rules and visible light feature parameter to achieve the visible light imaging feature matching aimed at the visible light imaging feature which improved the disposal speed and exact rate of recognition;(3) This paper discussed the multi-subtemplate correlation algorithm, and the correlation effect of each subtemplate is analyzed synthetically. Based on the analysis, a suitable template updating strategy is given, it can be adjusted automatically based on the change of the tracking object and the object is tracked stably. We also presents a Kalman filter to estimate state parameter in realize target tracking. And the algorithm uses the forecast function to reduce search areas for model matching. The experiments show the solution decreases the complex of correlation tracking, and has the advantages of good accuracy and high speed as well;(4) Aimed at the real-time demand of the system, the paper has presented a primary implemented scheme based on DSP and FPGA technology. The system include four portions: image gathering module, image pre-processing module, image processing module, image displaying module. The disposal speed is improved because of making...