Research On ISAR Imaging And Image Enhancement Methods Under Low SNR

Inverse synthetic aperture radar(ISAR)has the advantages of all-day,all-weather,longdistance observation and high resolution two-dimensional(2D)imaging ability of the noncooperative targets.It has been widely used in both military and civilian fields.The basic methodologies of ISAR imaging have been well developed in recent decades,and wellfocused imaging results can be obtained under high signal-to-noise ratio(SNR).However,with the increase of the ISAR observation distance,the SNR of the ISAR echoes will decrease rapidly,resulting that the traditional translational motion compensation method cannot achieve good performance.Hence,the imaging results of the observation target will be defocused.At the same time,the visualization performance of ISAR images under low SNR is so poor that it is difficult to accurately analyze the morphological structure of the observation target.In view of the above problems,this thesis mainly studies the joint translational motion compensation method based on non-parametric Bayesian inference theory and ISAR image enhancement method under low SNR.Accurate estimation and compensation of the translational motion can be achieved,and the visualization of ISAR image can be improved.The research content of this thesis mainly includes the following three parts:1.The basic theory of ISAR imaging is introduced.Firstly,based on the turntable model,the target echo signal model of ISAR imaging is derived.Then,the flowchart of ISAR rangeDoppler imaging algorithm is given,including the pulse compression,translational motion compensation and migration through range cells correction.This chapter provides the basis for the following research on joint translational motion compensation and image enhancement methods.2.To tackle with the problem that traditional ISAR translational motion compensation methods cannot achieve accurate estimation and compensation of translational motion accurately under low SNR,a nonparametric Bayesian translational motion compensation and high-resolution imaging methods are proposed.Firstly,since the coherent accumulation time is usually short in ISAR imaging,the translational motion of the observed target is modeled as a high-order polynomial form.Then,the prior probability distribution models of the scatterer and the translational motion are designed,and the nonparametric Bayesian posterior inference model for the estimation of the scatterer and translational motion is constructed.After that,the reverse jump Markov chain Monte Carlo algorithm is utilized to achieve accurate estimation of the translational motion.Finally,the compensation phase is constructed for the translational motion compensation,after which focused imaging result can be obtained.The effectiveness and robustness of the proposed method is verified via the experiments based on the electromagnetic data.3.To tackle with the problem that the poor visualization of the ISAR image cannot characterize the morphological structure of the target under low SNR,the ISAR image denoising and enhancement methods are proposed based on the morphological processing and retinex theory.Firstly,to achieve the ISAR image denoising under low SNR,the ISAR image denoising methods based on wavelet transform and morphological processing are studied.Then,to make the weak components of the target in ISAR images clearer,histogram equalization and retinex theory-based ISAR image enhancement methods are studied to improve the visualization degree of ISAR images under low SNR.At the same time,several ISAR image quality evaluation criteria are utilized to evaluate the performance of the image denoising and enhancement.Finally,the effectiveness and robustness of the proposed methods are verified based on the real imaging results of TIRA system.