Automatic Target Recognition And Tracking Technology For The Two-Color IR Imaging Guidence

The two color IR imaging guidance information processing is the key technology of the two color IR imaging seeker researching and manufacturing. The paper according to the practical application and the technological demand of the Two Color IR Imaging Seeker Image Processing System, researched the techniques of target detection,recognition and tracking based on the IR dual band image sequences comprehensively.The paper totally has five chapters.The technology of two color IR imaging guidance information processing is reviewed comprehensively in chapter one. Aim at the actual demand of the two color IR imaging seeker image processing system, proposed the functional frame of the two color imaging guidance information processing based on the analyzing and comparing of these schemes for the precision guidance image information processing in hometown and foreign literatures, and the primary researching contents of the paper and the key problems to resolve.In chapter two, aim at the problem of weak target detection under the highly cluttered background, proposed a method of weak target based the IR dual band image information fusion. This method according to the basic idea of algorithm fusion and information fusion, take advantage of the complementary and redundant information contained in the IR dual band images at the different stage of target detection, improved the performance of the whole system's for IR image weak target detection to a great degree.In chapter three, proposed a method of the IR dual band image dot target fusion recognition based on multi-features and hierarchical classification to resolve the problem of dot target recognition under the complex condition. The method firstly extracted the multiple features of target include the image features and temporal features, then classified these candidate target sub-image areas according to the feature vector representation using the adaptive K-Nearest Neighboring rule based the D-S evidential theory and the class prototype vector collection hierarchically. The method improved the system's target recognition efficiency to a great degree, presented a appropriate technique for the IR target and decoy discrimination.In chapter four, aim at the problem of maneuvering target tracking in the two color IR imaging system, proposed an algorithm of the two IR imaging target precision tracking using fuzzy inference adaptive weighting fusion. This algorithm firstly estimated the offset of target in the two different band IR images by processing the real-time target images form the middle wave and long wave IR sensors respectively using the proposed method of IR imaging target based on algorithm fusion. Then adaptive weighted the local offset estimations of the two IR sensors using the fuzzy inference technique, and filtered and predicted the offset of target in IR image using the least square error adaptive lattice filter. The algorithm take advantage of the superiority and data redundancy of multiple sensors system in the application area of target states estimation, improved the target tracking precision and robustness of the whole system and showed the preferable application perspective.Chapter five summarized the research content and primary contribution of this paper and probed into the application possibility of the two color IR imaging technology in these relative area.