Super-resolution Imaging Algorithms For Synthetic Aperture Radar (sar) Study

Synthetic aperture radar (SAR) is a kind of active coherent high resolution imaging radar system capable of producing high-resolution terrain image from data collected by a physically small aperture antenna. SAR achieve high angular resolution by integrating the backscattered signals to synthesize the effect of a large aperture antenna, and get the high enough range resolution by pulse compressing techniques.The image of the illuminated scene reconstructed from the collected two-dimensional SAR raw data is an inverse problem. In essence, the reconstruction is a parameter estimating problem, i.e., estimating the coordinates of the object in the slant plane and the radar cross section (RCS) of the objects. The classical imaging algorithms are based on the matched filter and get the images through processing the received raw data in frequency domain by broadly using the fast Fourier transform (FFT). Although the classical algorithms are proved very effective, robust, and can satisfy image quality in some applications, the conventional ones still have some shortcomings and limitations in principle. For example, it fails to achieve maximum SNR of the output when the noise in the input data sample is not additive gaussian stochastics. And its destination isn't to satisfy the fidelity of the output, but to achieve the maximum output SNR. So matched filter imaging algorithms improvements and imaging algorithms by other ways appeared. At first, in this thesis the rationale of synthetic aperture radar imaging is briefly synopsized. The reconstruction of the simulated point targets and actual targets in the scene are also done by the simulated and raw data, according to the imaging rationale. Then several auto-focus algorithms are discussed in detail, especially their performance in estimating the phase-error function, and followed by several numerical experiments. Because of the importance of speckle reduction in SAR image processing, the speckle reduction algorithms are categorized and compared, also with numerical experiments.Further in this thesis, the super-resolution imaging algorithms are classified according to their principles and analyzed briefly. And make some numerical experiments to testify some algorithms by simulated and actual raw data.