Electromagnetic Simulation SAR Imaging Under The Framework Of Compressed Sensing

Synthetic aperture radar (SAR) can obtain high resolution of range direction byusing pulse compression, and acquire high resolution of cross-range by using virtualantenna array formed by the motion of radar. Thus two-dimension image with highresolution for large scene is obtained. Because the method for actual measurementstakes not only a large amount of time but also a large amount of money to get the imageof target, it is necessary to research the image formation methods based on theelectromagnetic simulation. However, the simulation time is proportional to the datapoints required in imaging when we simulate the target echo signal. For general areatarget model, the simulation is extremely time consuming. However, compressedsensing (CS) algorithm can reconstruct the image by using fewer data points,significantly reducing the time it takes for the electromagnetic simulation.This paper can be summarized as follows:(1) The theory of electromagnetic scattering of complex target, the workingprinciple of the SAR and CS are analyzed, which lay the theoretical foundation for theresearch of electromagnetic simulation under the framework of CS theory.(2) The thesis introduces the basic principle of turntable imaging and fundamentalformula of the imaging. And the establishment of platform of SAR electromagneticsimulation imaging is mainly focused on. The research of the established simulationplatform includes two aspects: selection of computational electromagnetic method andthe setting of electromagnetic simulation parameters. In order to verify the effectivenessof the simulation platform, the image of the target with high resolution in simpleenvironment is formed. Then in order to simulate SAR imaging in real scenes, the targetin complex scene is imaged. First of all, the complicated backgrounds are constructedthrough the modeling of the rough ground surface, and then the target in the backgroundof rough ground is imaged. So the effectiveness of the platform is further verified.(3) Traditional electromagnetic simulation methods need to satisfy the Nyquistsampling theory, so all data points in the range and cross-range directions should besimulated. In allusion to the requirement of great amount of data, CS is combined with SAR imaging. Thus the model of SAR imaging based on CS is constructed and appliedin image formation of simple environment. In this way, the image formation problemconverts into the reconstruction of sparse signals. So it is possible to use CS toreconstruct original figure with fewer data points.(4) The basis pursuit (BP) algorithm and orthogonal matching pursuit (OMP)algorithm are used to image. The BP algorithm reconstructs the image with normminimization, and OMP algorithm constructs sparse approximation by looking for theatoms which match the measurement signal from the sensing matrix. The simulationresults show that both of the two algorithms can reconstruct the original image withsuper resolution by using fewer data points. Thus the electromagnetic simulation time isapparently reduced.