Studies On SAR Autofocus And Complex-valued One-bit SAR Fast Imaging

Synthetic aperture radar(SAR)is one kind of high-resolution imaging radar that plays an important role in the military and civilian fields.The classic Nyquist sampling theorem shows that,to obtain high-resolution SAR images,a large amount of echo data needs to be collected,which will impose great challenges to the data storage and transmission at the receiver.The emergence of Compressive Sensing(CS)not only effectively solves the problem of excessive data collection,but also further improves the imaging resolution.Therefore,SAR imaging based on CS has received extensive attention.However,CS reconstruction algorithms have large time and space complexity,and thus face tremendous challenges in practical applications.Moreover,one-bit CS reconstruction algorithms are often only applicable to the real number domain,which cannot directly apply to SAR without some manipulation beforehand due to the complex-valued data.In addition,due to the motion error of the radar platform,the observation matrix will introduce uncertainty,which will eventually appear as the phase error in the echo data,thereby defocusing the reconstructed scene.Therefore,in practical applications,CS-SAR auto-focusing algorithm becomes particularly important.Auto-focusing algorithms based on CS-SAR have been extensively studied.Such algorithms can correct small phase errors,but they are not capable of correcting larger error phases.This thesis investigates the insufficiency of existing CS-SAR auto-focusing algorithms and the low efficiency of existing one-bit CS-SAR imaging algorithms.The thesis is organized as follows:The first part of our research is entitled "SAR autofocus algorithm based on phase-free transformation".First,the shortcomings of the existing CS-SAR autofocus algorithm are analyzed,and then a specific phase error model is given.Finally,a convex model is formulated via phase-free transformation,so that the algorithm can reconstruct accurately focused images in the case of error phases of any size.The second part of our research is SAR autofocus method under one-bit compressive sensing framework.First,the thesis analyzes the challenges imposed by the one-bit CS-SAR auto-focusing technology,and then extends the phase-free transform methodology to the one-bit CS-SAR error phase correction problem.Combining the characteristics of the one-bit compressed sensing,a convex model is presented accordingly.Finally,simulation verified the effectiveness of the proposed method.The third part of our research is entitled "complex single-bit compressed sensing SAR imaging algorithm based on inverse chirp scaling operator".The algorithm replaces the exact observation matrix with the inverse chirp scaling operator,which greatly reduces the time and space complexity of the algorithm.Moreover,since the objective function based on the least square sense is adopted,the algorithm can be directly applied to the complex-valued data.Simulation and experiments both verify the effectiveness of the algorithm.