Studies On Compressive Sensing SAR Imaging And Autofocusing Technique

Synthetic aperture radar (SAR) could form a large equivalent aperture antenna to obtain highazimuth resolution imageries by signal processing techniques. Because SAR is a coherent imagingsystem, the received signal will suffer from a serious phase error, Because of the non-ideal motion ofthe aircraft. In order to obtain high resolution SAR images, the autofocus techniques are necessary toestimate and correct the phase error. Compressive sensing(CS) theory, which is proposed in recentyears, has a significant effect on acquiring and transferring a super vast amount of data. Theincoherent measurement in CS theory can solve the acquiring and transferring problem in highresolution SAR imaging effectively. The application of the CS theory in SAR imaging will induceimportant changes in the field of the radar technique. This paper will study on the imaging andautofocusing algorithm based on the CS theory. The implementation of the imaging and autofocusingmethod based on the CS theory is analyzed, the algorithms are proposed, and the experimentsincluding both point target simulation and field data indicate the effectiveness of the algorithmsinvestigated in this paper. This thesis are organized as following：Chapter1is the introduction of the thesis. The history of the SAR, SAR based on CS imaging,and autofocusing technology are outlined, and their development is introduced. At the end of thechapter, the aim and contents of our work are addressed.Chapter2introduces the CS theory. First, the theory is stated, and several common measurementmatrices and base matrices are recommended. Then the method of constructing sparse echo signal isanalyzed according to the characteristics of the SAR image data. Appropriate sparse basis andmeasurement matrix are chosen. After that, some reconstruction algorithms are outlined, and theorthogonal matching pursuit (OMP) algorithm is chosen accommodating the efficiency and theaccuracy of the algorithms.Chapter3studies the application of the CS theory in SAR imaging. The Range DopplerAlgorithm(RDA) and Polar Format Algorithm(PFA) are introduced first, among which the PFA usesthe scaling approach in both range and azimuth directions. Then an azimuth sparse signal is formedaccording to the discussion in the former chapter, and a point target simulation is performed usingboth RDA and PFA aiming at complete signal and sparse signal. The result shows that both PFA andRDA are suitable for sparse signal, while RDA performs not so good because of the absence of therange cell migration correction(RCMC) step. Chapter4studies the application of the CS theory in SAR autofocusing. First, the autofocustheory is introduced, and the generation and the types of phase error are analyzed. Two of theautofocus algorithms are recommended, and an autofocus algorithm using PAST is proposed, whichcan reduce the huge amount of calculation while using the MLE algorithm. This PAST algorithm haslow complexity and high accuracy. Then the improvement measure aiming at the sparse signal isdiscussed. At last, experimental data experiment is proceeded to validate the performance of thealgorithms and to compare the merits.Chapter5summarizes the thesis and points out the direction of research in the future.