Research On Imaging And Motion Compensation Method Of Linear Array SAR

Linear Array SAR(LASAR)is an emerging microwave 3-D imaging technology,has the ability of spatial 3-D resolution and provides richer information.Compared with the traditional SAR,its imaging model is more flexible such as downward-looking,forward-looking,side-looking and oblique-looking,etc.When working in the downward-looking model,the shadow-effect which exists in the traditional SAR can been overcame,so LASAR has broad application prospects in the various fields of military and civilian production.At present,the research of motion compensation problem on LASAR is still in the initial stage,and the imaging algorithm is inefficient.Therefore,this thesis mainly researches the motion compensation method and imaging algorithm of LASAR based on the backprojection(BP)algorithm.The specific contents and contributions are as follows:1.This thesis first introduces the basic theory of LASAR.The working principle of LASAR is described and the echo signal model is given.From the system fuzzy function,the theoretical spatial resolution is given.The 3-D RD algorithm and the 3-D BP algorithm are introduced,compared with the RD algorithm,the BP algorithm has high resolution and is suitable for any trajectory and any imaging model.Therefore,this thesis uses BP algorithm as the basic imaging algorithm.Finally,the basic principle of autofocus algorithm is introduced,which provides the theoretical basis for the following study.2.The motion compensation method of LASAR is researched.Firstly,the error sources of LASAR are analyzed.Based on the traditional autofocus BP algorithm,a 3-D autofocus BP algorithm based on phase error estimation is obtained by extending the 2-D to 3-D.This algorithm models the motion error as the phase error in the echo and uses the optimization method with image sharpness maximization criterion to estimate the phase error.The simulation results show that the algorithm has good focusing performance for all kinds of phase errors,and the effectiveness of the algorithm is verified by processing the experimental data.However,because the method assumes that the phase error of all the pixels in the scene is same,the spatial variability of the motion error is ignored,this can lead to decrease in the estimated accuracy when imaging a large scene.For this problem,a 3-D autofocus BP algorithm based on APC error estimation is proposed.The algorithm estimates the APC measurement error with the image intensity maximization criterion.Since estimating APC is equivalent to estimating the range history of each pixel,this method can perform high precision for phase compensation of each pixel.At the same time,this method greatly reduces the demand for processor memory and has the advantage in solving the LASAR motion compensation problem in the large scene.Simulation and experimental results verify the effectiveness of the proposed algorithm.3.A 3-D frequency-domain backprojection(FDBP)imaging algorithm is researched.This method forms 3-D SAR images by backprojection the spectrum of a radar echo into the image wavenumber domain to obtain the spectrum of the image,and then transforms the spectrum of the image into the image domain by inverse Fourier transform to obtain 3-D SAR image.Compared with time-domain backprojection(TDBP)algorithm,FDBP algorithm not only has the advantages of high accuracy,but also has the advantages of higher imaging efficiency.The simulation results show that FDBP has more than ten times efficiency improvement compared with the TDBP algorithm.4.A method of parallel implementation of 3-D FDBP algorithm based on GPU is proposed.Using the parallel structure of FDBP algorithm,it is implemented on GPU platform with pixel-based parallelization scheme.Compared with the CPU-based FDBP algorithm,there is a great increase in efficiency,and the bigger the scene,the more obvious the acceleration ratio,so it has the advantages in engineering applications.