Research On Plane IR Imaging Detection, Recognition And Tracking Technology

With the rapid development of IR imaging and its widest prospect in military application, the IR image processing technology has been the research hotspot in relative professional field. According to the current development status of IR image processing methods, and the main problems existed as well as the basic requirements on its practical engineering applications, this paper has made systematic studies on various aspects such as detection of IR images, target recognition and target tracking. It has analyzed and compared with the typical arithmetic nowadays, and the experimental results have been presented in the paper, too. Considering of the IR imaging characteristic of subject and its background, some effective methods are presented and verified by the computer's software. The main achievements of this paper are as follows:(1) A new edge segmentation method based on multi-scale wavelet transformation thinking of the target's size in the scales which size is gained by the Kalmen forecast model is presented. So that the self-adaptability of wavelet transformation method is improved, then the system's effective detection and segmentation ability is improved in the complicated background of the war.(2) Aimed at the feature of research subject and the disposal results of wavelet transformation, a segmentation method in which the mean of imaging gray-degree is considered as partition value is presented based on the difference image of wavelet transformation. And the intersected effect is compared with the mean method and the best partition value of basal and entire segmentation method, and it is verified that the method is simple and effective, through which not only the computing speed is improved, but the disposal effective is not influenced much.(3) At the aspect ol^ feature matching, the author put forward the feature vector that amalgamated the changeless rules and IR feature parameter to achieve the IR imaging feature matching aimed at the IR imaging feature which improved the disposal speed and exact rate of recognition.(4) Aimed at the real-time demand of the system, author has presented an implemented scheme based on DSP and FPGA technology. The disposal speed is improved because of making the DSPs CMOS chip cooperating with FPGA chip as system's central CPU. And adopting the FIFO component and EDMA disposal technology ensure the system's real-time disposal ability.