3D Electromagnetic Feature Extraction Of High-Resolution SAR Targets Based On Parametric Scattering Model

High-resolution Synthetic Aperture Radar?SAR?three-dimensional?3D?image contains rich dentifying features,which has become a research hotspot in radar target recognition.The content of this paper is mainly focused on SAR 3D image reconstruction method and 3D electromagnetic scattering feature extraction method.The basic electromagnetic feature of the high resolution SAR image is the scattering center.In this paper,we first analyze the modeling process of bistatic 3D scattering center,and verify the validity of the 3D attribute scattering center model by using the FEKO high-frequency electromagnetic calculation software.Then aiming at the problem of 3D reconstruction of high resolution SAR image target,based on two-dimensional Fourier traditional imaging methods,analysis the interpolation process of 3D data,the Fourier 3D reconstruction method based on nearest neighbor interpolation is presented.Also,the method of 3D electromagnetic feature extraction based on image domain is studied,the parameter estimation and structure selection process in feature extraction are selective analyzed.Finally,the validity of the electromagnetic feature extraction method based on Fourier 3D reconstruction is verified by the FEKO simulation data.The three-dimensional parameterized scattering center model can well characterize the electromagnetic characteristics of the target.First of all,the establishment of bistatic three-dimensional scattering center model is analyzed,and the validity of the model is verified by the FEKO high-frequency electromagnetic computing software.Then the Fourier 3D imaging method and the electromagnetic feature extraction process based on the image domain are studied.Finally,based on the FEKO simulation data,the validity of the 3D electromagnetic feature extraction method based on the image domain is verified.However,in fact,the three-dimensional data of high resolution SAR images is often sparse.The traditional Fourier imaging method will bring some problems such as poor reconstruction quality,large memory and long time consumption.Based on the compressed sensing theory,this paper presents an improved smoothing l₀ norm?SL0?sparse reconstruction algorithm.In terms of image quality and computing efficiency,Fourier reconstruction method,l₁ norm minimum reconstruction method and the proposed algorithm are compared respectively,which proves the superiority of the algorithm proposed in this paper.Then,based on sparse reconstruction image,we extract the 3D attribute scattering center feature of simple target scatterers,and prove that the feature extraction method based on sparse reconstruction improves the accuracy of feature extraction.In order to expand the application of 3D electromagnetic feature extraction method based on sparse reconstruction,Aiming at the problems of complex aperture,such as wide aperture,large volume and scattering anisotropy,a sparse reconstruction method based on incoherent combination of subapertures is proposed.The method divides the wide aperture into multiple subapertures,and uses the improved SL0 algorithm proposed in this paper to sparse reconstruct,then incoherently combines the sub images.Finally,the experiment of electromagnetic feature extraction based on sparse reconstruction is carried out on the 3D SAR data of the backhoe,which validates the effectiveness of the proposed method.