期刊名称：International Journal on Smart Sensing and Intelligent Systems
印刷版ISSN：1178-5608
出版年度：2015
卷号：8
期号：1
页码：497-515
出版社：Massey University
摘要：This paper studies the multi-resolution analysis of infrared image preprocessing methodbased on wavelet transform, wavelet transform infrared image of small target for pretreatment, afterpretreatment suppressed image background clutter, improved signal to noise ratio. On this basis,studies based on infrared image sequences generated background Kalman filter and target detectionalgorithm, given the principle of the algorithm, time-domain method for generating fading backgroundand moving object extraction methods such as recursive least squares method, and the experimentalresults show that the algorithm capable of detecting small targets in infrared images, with good results.based pretreatment small target motion continuity characteristics studied image sequences smallmoving target detection algorithm based on weighted dynamic programming, dynamic programmingalgorithm principle is given, analysis of the direct method and the accumulation of gray dynamicprogramming algorithm based on similar likelihood function for dynamic programming problems ariseenergy diffusion gives improved dynamic programming algorithm: weighted dynamic programmingalgorithm, the algorithm implementation steps, testing and experimental structure the results wereanalyzed, the results show that the algorithm for small moving target detection with good results.Finally, this paper studied the infrared image reject false target trajectory get the real target trajectorycorrelation method, so as to improve the detection rate and reduce the false alarm rate the purpose.
