Research On Image Formation For General Bistatic SAR

As a new concept SAR which separates transmitting and receiving stations,comparing with traditional monostatic SAR, bistatic SAR system propose some obviousadvantages, such as improving object scattering factor, increasing system survival andenhancing antijam capability etc. According to the universal law of development,general bistatic SAR with different velocities of transmitter and receiver, is the headingof development due to its flexibilities. However, the flexible configuration leads tosome difficulties for SAR image formation, which are the key research projects in SARprocessing.Sponsored by the Natural Science Foundation of China, This dissertation focuseson image formation for general bistatic SAR.In the aspect of imaging theory,"Equivalent Monostatic Range History" isproposed to use traditional monostatic SAR image algorithm to focus bistatic SAR data."Point Targets Set with Same Range History Function" is builded to analyse the space-variant feature in range history of bistatic SAR. According to matched filtering, themathematical model of image formation is builded for designing image formationalgorithms in bistatic SAR. Based on range and Doppler, the spatial resolution isderived to evaluate the performance of bistatic SAR system and to design the flightscheme.In the aspect of image formation, this dissertation proposes four bistatic SARimaging algorithm though different set of eyes, which are validated by airborne bistaticSAR point target echo data, satellite-airborne bistatic SAR point target echo data andairborne bistaitc SAR raw data.1) A bistaitc SAR polar format algorithm based on two-dimensional nonuniformfast Fourier transform is proposed. According to matched filtering principle, bistaticSAR polar format algorithm is derived. To improve the utilization ratio of echo data inpolar format using traditional interpolation method, two-dimensional nonuniform fastFourier transform is derived to solve the nonuniform data mapping problem. Thus,bistatic SAR polar format algorithm based on two-dimensional nonuniform fast Fourier transform is proposed, which can make the best of echo data and obtain ideal focusingperformance.2) A bistatic SAR image formation algorithm based on generalized keystonetransform is proposed. To improve the lack of keystone transform which only canhandle the linear range cell migration, a novel data transformation named "generalizedkeystone transform" is derived to correct the nonlinear range cell migartion. Thus,bistatic SAR image formation algorithm based on generalized keystone transform isproposed, which can obtain better focusing results.3) A bistatic SAR image formation algorithm based on optimal image space isproposed. Considering the mean-square-error of the truncation range error as thecriterion for evaluating image space, the optimal image space in mean-square-errorsense is derived for bistatic SAR according to Karhunen–Loeve theorem. Thus, bistaticSAR image formation algorithm based on optimal image space is proposed. Theminimum mean-square-error can be obtained using this optimal image space, and cangenerate the ideal focusing performance.4) A bistatic SAR frequency domain image formation algorithm based on MSRspectrum is proposed. For the lack of geometric sense in the MSR frequencey spectrum,a modified expression is derived to design frequency domain imaging algorithm.According to the traditional Omega-K principle, bistaitc SAR echo data can be focusedappropriately.In the aspect of bistatic system error compenstation, this dissertation summarizesthe platform motion error, time synchronization error and frequency synchronizationerror as the phase error, and builds the integrated signal model with phase error. Basedon this integrated signal model, the phase error is divided into two parts: fast frequencyindependent phase and fast frequency linear dependent phase. The standard of phaseerror compenstation is analysed under four error forms. The corresponding phasecompenstation method is proposed, which is based on CUDA parallel processing.In conclusion, this dissertation does deeply research on imaging theory and methodfor bistatic SAR. The research results are useful for designing bistatic SAR system andsignal processing.