Study On SAR Data Processing Algorithms And Data Processing System Based On Clusters

This thesis focuses on data processing technologies of Synthetic Aperture Radar (SAR), and the SAR data processing system based on clusters. The purpose of the thesis is studying and improving the prior SAR data processing algorithms, and establishing a new SAR data processing system, with high processing speed, high quality images and high performance/cost ratio.First, the thesis studies the SAR data processing algorithms. In the study on SAR imaging, this paper studies and analyzes the effect of phase error of range and azimuth directions on SAR imaging. And with the practical SAR data, the effect of range phase error compensation method used by this paper is proved. In the post processing of SAR images, the paper studies the partial differential equation, and establishes a new diffusion coefficient both on the local statistics of SAR images and the robust estimation. The performance of the new filter is proved in the experiments,In the interferometric SAR processing, according to the problems of the prior phase filters, this paper proposes a new phase filter based on the anisotropic diffusion and gets good results.Second, this paper studies and implements a SAR data processing system. Based on the introduction of the parallel and distributed processing, this paper analyzes the parallel partitioning of both SAR data and SAR data processing method, and compares t he p erformance o f t he d ifferent p artitioning m ethods. At last, t his p aper designs the construction, components and messaging driving mechanism, and describes the implements of the main modules.The SAR data processing system has processed stably a large amount of data in many practical aviation tasks of the Institute of Electronics of Chinese Academy of Sciences, and the system has many advantages, such as producing high quality images,...