Research On The Algorithm Of Ship Detection In SAR Image

Ship Detection in SAR Image is of great significance for marine surveillance and fishery management.In civil use,the detection of sea areas can help search and rescue.In terms of military affairs,it can achieve traffic control and detection of illegal ships of other countries.Two aspects are mainly concerned in this thesis: clutter distribution identification and ship target CFAR detection algorithm.Firstly,the features of SAR images and the characteristics of ship targets are introduced,and the main factor that affect the detection of ship targets is analyzed.On this basis,the theory of ship detection on CFAR is studied,and the structure and detection performance of multiple CFAR detectors are analyzed to provide reference for the design of ship target detection algorithm.Secondly,common sea clutter models are studied,and the parameter estimation methods of clutter model are given.The applicable scope,advantages and disadvantages of different clutter statistical models are summarized.On this bases,the hypothesis tests of Rayleigh,Weibull,Lognormal and K distribution are implemented by introducing the Anderson-Darling goodness-of-fit test method.Combined with Anderson-Darling test and parameter estimation,an algorithm of clutter distribution identification is proposed.The validity of this method is proved by simulating clutter identification and real clutter distribution models identification.Finally,a TLM-CFAR detector is proposed to solve the tedious problem of parameter estimation of MLH-CFAR detector.The detector uses TL moment estimation to obtain Weibull two parameters.At the same time,in order to solve the multi-objective interference problem of Weibull distribution,a cycle elimination algorithm was added to obtain the cycle elimination TLM-CFAR detector.Performance analysis results and test results prove that the detector has good detection performance.