Research On Infrared Detection And Tracking Tech Of Small Target For Complex Interference Environment

Based on missile signal processing, this paper mainly studies in infrared target detection, tracking and infrared automatic target recognition.Adaptive background estimation is resershed according to the character of infrared image. The mathematic model of adaptive filter is deduced and the formular that is used in engineering is given out. A method of adaptive background estimation based on gaussion core aimed at the defect of adapative filter. This method is applied in wavelet coefficient to detect target.An infrared imageing object tracking algorithm is bringed forward aimed at complicated background object segmation. First, the method divisions the input images using canny edge detection algorithm. second, the method matched the scene and compute the scene parameter. Third, differential processed to the result according to this parameter. In the end the method binarizied the result based on part entropy segemation.Aimed at the small target tracking problem, a method of multidimensional assignment algorithm based on Particle Swarm Optimization is proposed in the base of the results of the static problem of associating measurements. Through the problem of finding the minimize solution of the objective cost function in multiple dimensional assignment is express as a combinational optimal problem. Furthermore, the best seclusion can be found fast in respect of reducing the search range through adding measure restriction.The paper first generalized the yager fusion formula and a better fusion formular is givenout. The new fusion formular is better than D-S evidence theory formula and yager formula.A method of IR Small target fusion recognition is proposed based on the D-S evidence theory in this paper. This method firstly distilled small target character from the far infrared and middle infrared image, and then determined the basic mass function based on object character, in the end identified the target based on the D-S evidence theory. This paper presents a model of optimized fusion based on consensus theory and fuzzy reasoning for attribute classification. According to the distribution of consistency and reliability in classification decision set, the intelligent fusion model can analyze and judge sensor information based on fuzzy reasoning, and distinguish different information conflict, and decide corresponding fusion strategy. Finally, classification decisions are fused based on the evidence characteristics. The fusion scheme ensures that conflicting decision evidences can be fused effectively, and makes reasoning fusion of high level information more intelligent and have actual meaning.