Research On Imaging Theory And Key Technology Of Video Synthetic Aperture Radar

The research on video synthetic aperture radar(ViSAR)imaging theory and its related applications is one of the most advanced topics in the current field of SAR technology.ViSAR can form continuous frame images with high frame rate and extend the static scattering information of traditional SAR image to dynamic scattering information,providing a new way to conduct SAR post-processing from the temporal dimension.Among various applications of ViSAR,ViSAR shadow-based moving target indication(MTI)is one of the research focuses.By combining the advantages of shadow-based MTI and ViSAR shadow imaging,this technology can effectively avoid the problems of traditional MTI technology,and improve the detection probability,positioning accuracy,and real-time capability.However,all ViSAR applications are built on the premise of real-time high-resolution imaging,which brings forward higher requirements on both the efficiency and accuracy of ViSAR imaging algorithm.Based on this,this thesis has studied the theory and key technologies of ViSAR imaging,and conducted an in-depth discussion on high-frame-rate imaging,multi-frame imaging,real-time imaging,and curvilinear-mode imaging,hoping to contribute to the development of ViSAR system in China.The main content and contribution of the thesis include the following aspects:1.Based on the principle of SAR imaging,the factors affecting SAR frame rate were analyzed,and the definition and comparison of high-frequency ViSAR and non-high-frequency ViSAR were made.The concept of ViSAR persistent imaging was proposed for the demand of persistent ground monitoring,and the Starring Spotlight SAR mode and persistent imaging processing of ViSAR were introduced.As for the research hotspot of ViSAR shadow-based MTI,both the advantages of shadow in MTI and the advantages of ViSAR in shadow imaging were analyzed.Finally,the basic theory of ViSAR persistent imaging was studied and the feasible imaging algorithms were analyzed and compared.2.In order to solve the problem of persistent imaging in circular ViSAR,several circular PP algorithms were proposed.Firstly,the circular ViSAR persistent imaging problem was decomposed into coordinate system matching and resolution matching,and the matching conditions were analyzed and derived.Based on this,two kinds of circular PP algorithms based on Fourier transform were proposed,namely GPP algorithm and LPP algorithm.In addition,by combining coordinate system matching with Fourier transform imaging,circular persistent imaging was realized by a single 2D-CZT,and the CPP algorithm was proposed.The computation burden of GPP,LPP,and CPP was analyzed and compared.For the angle-dependent scattering problem in wide-angle SAR,an angle-dependent scattering imaging algorithm,namely AGLRT algorithm,was proposed based on the high frame rate characteristics of ViSAR.Finally,the validity of the above algorithms was verified by both numerical simulation and measured data.3.The problem of circular PP frame image distortion for close range or large scene was studied,and several circular EPP algorithms were proposed.Through analyzing the relationship between the approximate slant range under the plane-wave assumption and the actual slant range,the relationship between the targets' reconstructed position by circular PP algorithm and their actual position was derived,namely circular PPM.Based on circular PPM,through interpolation-based correction of the circular PP frame images,persistent imaging with distortion correction was realized,LEPP,GEPP and CEPP were proposed.In addition,by combining Fourier transform imaging with circular PPM,the process of imaging followed by correction was realized directly by a single NuFFT-2,then NLEPP algorithm and NGEPP algorithm were proposed.Subsequently,the computational burden of all the above circular EPP algorithms were analyzed and compared,and the effective imaging scene extent of circular PP algorithm and circular EPP algorithm were derived.Finally,the validity of the above algorithms was verified by numerical simulation and their efficiency was compared.4.In order to expand the application conditions of ViSAR,persistent imaging in curvilinear ViSAR was studied,and several curvilinear PP algorithms were proposed.Firstly,the single-frame wavenumber spectrum in curvilinear ViSAR mode was analyzed,and an adaptive aperture division method based on resolution matching was proposed.On this basis,polar format algorithm was extended to curvilinear ViSAR persistent imaging,and FPP algorithm was proposed.Then,for the problem of imaging deterioration caused by the scatter interpolation error in FPP algorithm,N1 PP algorithm was proposed.In N1 PP algorithm,the non-uniform distribution of wavenumber spectrum was combined with Fourier transform imaging,where imaging was realized directly by a single NuFFT-1,avoiding the error and computational cost of the scatter interpolation in FPP.5.To solve the problem of curvilinear PP frame image distortion for close range or large scene,the circular PPM was extended to curvilinear mode,and curvilinear PPM was derived.Based on curvilinear PPM,through interpolation-based correction of FPP and N1 PP frame images,FEPP algorithm and N1 EPP algorithm were proposed,respectively.Furthermore,by combing the non-uniform distribution of wavenumber spectrum with the corresponding non-uniform imaging grids after curvilinear PPM,the process of imaging followed by correction was realized directly by a single NuFFT-3,and N3 EPP algorithm was proposed,avoiding all interpolations before or after imaging.Finally,the validity of the above algorithms was verified by numerical simulations and their efficiency was compared.