Study On SAR Fast Time-Domain Imaging And Motion Compensation

As a remote sensing detection technology,synthetic aperture radar(SAR)plays an important role in military target detection and civilian remote mapping.It has all-weather and all-time imaging capability.Efficient and fast imaging algorithms have always been the core content of SAR imaging technology,which need to be continuously studied and improved to meet various engineering applications.Frequency-domain imaging algorithms rely on twodimensional Fourier transform(FT)to compensate and correct the spectrum,which achieve full-scene imaging quickly.However,as the trajectory of radar platform changes and the squint angle of beam increases,the traditional frequency-domain imaging algorithms are difficult to achieve range-azimuth decoupling of echo data,resulting in poor focusing effect and failure in SAR image.Curve trajectory and large squint-angle imaging are also the difficulties to be overcome in SAR research.The time-domain imaging algorithms are based on point-to-point phase compensation and coherent accumulation to achieve imaging,and its non-approximation feature makes the time-domain algorithms suitable for any trajectory,any squint-angle view and any imaging coordinate system,which makes up for the defects of the frequency-domain imaging algorithms.However,the time-domain imaging algorithms are still faced with the problems,including insufficient applicability of fast time-domain imaging algorithm in multi-mode and lack of research on time-domain motion compensation,which undoubtedly limits the engineering application of time-domain imaging.Based on the National Natural Science Foundation of China and the horizontal projects,this dissertation conducts research on fast time-domain imaging and time-domain autofocus technology,and the purpose of this dissertation is to expand the engineering applications of time-domain imaging.The main research contents are as follows:1.Aiming at the problem of mismatching of fast time-domain imaging algorithms with arbitrary curve trajectory,the curve-trajectory fast time-domain imaging method is studied and generating mechanism of azimuth resolution is analyzed.The virtual standard linear aperture is constructed along the imaging azimuth direction.The instantaneous range of the target is calculated by the three-dimensional coordinate system.And the real aperture curve trajectory is projected onto the virtual aperture,so as to achieve fast time domain imaging under the curve trajectory without losing azimuth resolution.The effectiveness of the proposed method is verified by simulation.The proposed method can effectively expand the types of time-domain imaging platforms and meet the requirements of high-maneuveringplatform imaging.2.A two-stage time-domain autofocus method based on generalized sharpness and AFBP is proposed for the time-domain non-space-variant phase-error compensation.In this method,local data is used to replace the whole image,which improves the computing efficiency.The phase-error estimation problem is transformed into a global optimization problem with constant-modulus constraint,and the phase errors can be solved by genetic algorithm(GA).In addition,a maximizing-maximum-pixel-value(MMPV)method is proposed to further optimize the estimation problem.To solve the phase errors,the image sharpness is converted into the maximum value of space vector length by using the principle of conservation of image energy,which avoids searching and iterative process and greatly improves the efficiency of phase estimation.In the aperture fusion stage,the residual errors between subaperture are considered.And the residual errors are eliminated by linear search and image matching to achieve coherent image accumulation.Simulation and experimental data processing results show the effectiveness of the proposed method.Under the condition of non-spatial-variant phase errors,the proposed method can realize the time-domain image autofocus rapidly and improve the quality of image effectively.3.A time-domain autofocus method based on three-dimensional motion-error estimation is proposed to compensate the space-variant phase errors.The method uses the linear relationship between the multi-local phase errors and the three-dimensional motion errors,and realizes the correction of the radar platform trajectory by estimating the threedimensional motion errors.Then the point-by-point compensation is realized.In the process,the overlapping sub-block configuration is proposed to divide the whole image into multiple sub-blocks according to the distance direction.The residual phase errors are searched and corrected by image matching and coherent accumulation of the overlapping parts of subblocks.Considering the influence of phase winding on the estimation of real motion errors,a mixed integer programming model is established.A joint solution of Tikhonov regularization and GAis proposed to estimate the motion error.The effectiveness of the proposed method is verified by point-target simulation and measured data processing.The proposed method can improve the time-domain imaging quality of wide scene.4.Frequency modulated continuous wave(FMCW)and strip-map time-domain imaging are researched to expand the application of time-domain imaging.First of all,by analyzing the FMCW echo signal model,the range motion and residual phase are obtained after dechirp processing.By calculating real range and compensating residual phase,time-domain integral imaging is realized for FMCW.In strip-map time-domain imaging,the angle-overlapped unified polar coordinates is estimated to avoid redundant complementary angle-domain sampling and ensure the integrity of wavenumber spectrum.The full aperture of the strip is divided into standard apertures and sub-apertures.The full-aperture wavenumber spectrum is obtained by splicing the wavenumber spectrums of the sub-apertures.Then,the strip imaging is realized.The effectiveness of the proposed method is verified by point-target simulation and FMCW strip-map data processing.The proposed method can expand the mode of time-domain imaging and meet the needs of civilian mapping applications.