Study On The3D Imaging Of Manmade Targets Based On Polarimetric Radar

3-D reconstruction of man-made objects is an advanced issue in the area of polarimetricradar imaging, and is of great value both in theory and in applications. Aimed topromote the applications of polarimetric3-D imaging radar in military reconnaissance,this thesis studies the techniques including polarimetric SAR calibration and3-Dreconstruction of man-made objects in the background of polarimetric interferometricSAR (InSAR) and polarimetric tomographic SAR (TomoSAR).Chapter1illustrates the background and necessity of the research, briefly recallsthe development of polarimetric imaging radar systems, summarizes the previous worksin the related fields，and then gives a brief introduction of the major works in thisresearch.Chapter2studies the SAR polarimetric calibration techniques. After reviewingsome classical polarimetric calibration techniques based on point targets or distributedtargets, an in-depth performance study is presented through theoretical analysis andsimulation verifications. For the first time, it is found that there exist some definite butunreasonable relations between the estimates for the distortion parameters form theclassical Ainsworth calibration technique. For the case that the azimuth symmetryassumption widely adopted in the previous works does not hold, a calibration modelcalled TOP, which is capable of preserving the target orientation information, isproposed by considering the orientation angle shift effect on the measurement of targetscattering matrix. Then a fast iterative algorithm is proposed to solve the TOPcalibration model. In addition, the couple between the estimates for the orientation angleand the distortion parameters is studied, and the impacts of the error of orientation-angleestimate on the estimate for the distortion parameters are analyzed. Real collectedairborne SAR data are utilized to verify the effectiveness of the proposed model andalgorithm.Chapter3studies the3-D reconstruction of man-made objects using polarimetricInSAR. A processing scheme for the reconstruction of man-made objects of two kindsis presented firstly. Then, in-depth researches about two key topics, which are themulti-polarimetric joint3D imaging and the suppression of multiple scattererinterference, are presented. A super resolution technique (M-ESPRIT) for the jointextraction of scattering centers in polarimetric InSAR images is proposed, and theperformance of M-ESPRIT algorithm is checked by processing simulation data and theanechoic data. We derived the phase statistics for the interferometric phase of scatterers produced by the overlay of multiple elementary ones. A detection algorithm is proposedfor testing whether a scatterer is an elementary scatterer, and its performance versusSNR is presented. For the phase ambiguity problem, a phase alignment operation isrecommended. We demonstrate the effectiveness of the proposed techniques byperforming the3-D reconstruction of some targets using polarimetric InSAR data.Chapter4presents a thorough research on the3D tomographic imaging techniques.For the tomographic imaging in the case of irregular baseline distribution, twoalgorithms based on array transform technique and Tikhonov regularization technique,respectively, are proposed. We show from the statistical point that the Tikhonovregularization method is actually a MAP estimate for the height profile in someparticular cases. A united framework incorporating the FFT, TSVD and the Tikhonovregularization method is presented. For the first time, the distributed compressivesensing technique is introduced into the polarimetric SAR tomography application,producing the MMV-CS super resolution tomographic imaging technique. A fastimaging algorithm is proposed for the MMV-CS model, and the signal leakage problemis tackled with a window-based iterative algorithm. Simulations are performed to checkthe effectiveness of the proposed algorithms. The3-D full polarimetric tomographicimaging experiment is carried out. We demonstrate that the impact of the phase errorinduced by an array element error on the tomographic imaging, and propose acalibration technique employing prominent scatterers to adapt the3D imaging methodto practice. For the first time, the3D reconstruction results of a vehicle and cornerreflectors are obtained, and the results verify the advantages of the proposed MMV-CSmethod in terms of super resolution power and ambiguity level. Chapter5summarizesthe research work and main innovations, and discusses the future work.