Study On The Autofocus Algorithm For Circular SAR

Due to its intrinsic characteristics,circular synthetic aperture radar(CSAR)has recently become of particular interest to SAR community.Unlike Linear SAR mode,CSAR offers the potential to provide resolution of ?/ 4,obtain target information in all orientation and retrieve the power distribution of scatters in the height direction.Nevertheless,because of its special imaging geometry,most of traditional algorithms performed in frequency domain are not valid anymore,so,the time domain processing method is a better choice.Back Projection(BP)algorithm can precisely accommodate the complicated track of the platform,but the huge computational burden and the difficulty of combining with autofocus algorithm,especially the latter one,restrain its application severely.Due to the limited accuracy of Global Position System(GPS)in airborne SAR,it is usually impossible to obtain satisfied imaging results.So,effective autofocus algorithm is the key to acquire high quality image.In order to solve these obstacles,this paper focus on autofocus algorithm,the main study is as follows:1.This dissertation studies the fast domain processing algorithm for CSAR.The proposed method is based on the circular aperture of the CSAR model,and the full aperture is divided into several sub-apertures.Then,those sub-apertures are processed by the fast factorized back projection(FFBP)to get their corresponding sub-images.At last,the final focused image is obtained by the coherent interpolation of the sub-images from the polar coordinate to the global coordinate system.The proposed method has the advantages of high efficiency and large imaging area.2.Base on the Factorized Geometrical Autofocus algorithm of LSAR,this paper come up with the extend FGA algorithm for CSAR.The strategy is integrated in a fast factorized back-projection processing chain and relies on varying track parameters step by step to obtain the sharp image.To set up constrained problems,only one track parameter is corrupted in one simulation.3.This paper studies the Phase Gradient Autofocus approach for time domain method of CSAR.Firstly,based on the imaging geometry of CSAR,the expression of the motion error encountered in CSAR is presented.Then,on this basis,the signal formation of time domain processing is derived.Finally,the implementation of improved PGA is derived and presented in detail.In study procedure above,experiment results based on the simulated data and real data are given to prove the correctness of theory analysis and the validity of proposed method.