| A range distributed target detection in sea clutter is addressed in this paper.Because of the performance loss caused by the no widely applied single target in partially homogeneous radar environment,the group generalized likelihood ratio test(GLRT)developed here in partially homogeneous radar environment.In the estimated covariance matrix,for the phenomenon that the statistic property of sea clutter is equivalent to that of speckle ignores the influence of texture on the statistic property,a modified sample covariance matrix(MSCM)estimation based on the maximum a posteriori is proposed in the inverse gamma distributed texture.For the better estimating precision,the MSCM estimation is obtained by the inverse Gaussian distributed texture,and then applied to the GLRT based on the maximum a posteriori.Four parts are mainly included in the thesis:1)Firstly,it reviews the development history and research status of target detection,analyzes the mainly existing problems,and then summarizes the statistical characteristics and common models of sea clutter.2)Several classical target detection algorithms are analyzed briefly and then the group GLRT is developed here in partially homogeneous radar environment.In addition,the performance of the group GLRT detector and the previous detectors are analyzed through the simulated experiment,and the advantages and applicability are proved.3)Three estimators are briefly analyzed and then a modified sample covariance matrix estimator based on the maximum a posteriori is proposed for the inverse gamma distributed texture.Finally,the improvement of the detection performance for the proposed covariance matrix estimator is proved by the experiments.4)For the better estimation precision,the texture in MSCM estimator satisfies the inverse Gaussian distribution,and then is applied to the GLRT detector based on the maximum a posteriori for the inverse Gaussian distributedtexture.The good performance of the detector is proved by the experiments at last. |
PDF Full Text Request