The Application Of Wavelet Analysis In Infrared Target Detection

Since infrared imaging guidance is a very important guidance system with strongcapability to resist perturbations, it has become a main trend of development. Dim targetdetection and tracking is a crucial step in infrared imaging guidance. From faraway tonear, a target evolves from a point, then a unsteady speckle, and a face-shape target inthe end. The detection of targets in long distance with small spatial extent and lowcontrast in infrared clutter background is a problem of crucial importance, because itcan improve system sensibility to intruding targets and provide enough response time. Supported by the Preliminary Research Foundation of National Defense No.5140102021JW0506, this paper mainly studies infrared dim target detection by using ofthe powerful mathematic tool of wavelet analysis, further, it is mainly interested in thedim target detection end image enhancement with the ocean and sky background. Many problems are subsistent in infrared dim target detection and tracking, i.e.,dim targets have a small spatial extent and weakened morphological features; dimtargets often submerge in complex clutter background; many kinds of noises co-exist;sometimes targets can be occluded or temporarily lost for other factors. In view ofthis ,By means of drawing into the correlation theory of Shannon Information Tropic,we propose a infrared target detection algorithm, which acquire better effect ;Andthrough thoroughly analyses on the infrared image's local grey degree,we propose threekinds of infrared targets enhancement algorithm, Moreover, by utilizingseeded-region-growing algorithm,we achieve accurately collecting of the target 。 Being praised as "Mathematics microscope ", the wavelet analysis obtain thewide-ranging application in image processing ,pattern recognition and so on in the pastfew years。It is a perfect combination of function analysis, Fourier analysis, splineanalysis, harmonic analysis and numerical analysis, and been thought to be a greatbreak-through in theory and method in recent years. Combination the thoroughlyanalysis of the background,target and noise in the infrared image, we present aninfrared target detection method based on wavelet multi-scale analysis and Fishersegmentation after elaboration the wavelet transform and multi-scale. Above all, weproposed another infrared target detection method which is robust to noise through thecombination of the á trous algorithm and seeded-region-growing algorithm.