Research Of SAR Imaging Arithmetic Based On Compressed Sensing

With the rapid development of modern information technology, the signal sampling and processing is increasingly important. Since the high frequency signal sampling based on original Nyquist sampling theorem has a high demand of hardware system, a new technology which can break the traditional mode of sampling is urgently need to reduce the sampling frequency. Synthetic Aperture Radar (Synthetic Aperture Radar, SAR) is a kind of more frequently used high resolution Radar imaging technology. With the deepening of the research, a lot of high resolution imaging algorithms have been put forward and proved. But because of the Nyquist sampling theorem, all these algorithms are faced with high sampling rate, large amount of data, data transmission, storage, fast processing difficulties and other issues. As a new data acquisition and signal processing method,compressed sensing theory offers a possible way to solve the problem of high speed sampling.According to the above problem need to be solved, a new SAR imaging algorithm based on compressed sensing has been proposed to solve the high requirements for hardware because of the high frequency sampling in this paper, and it can be implemented through downsampling imaging process. In this paper, in view of two kinds of defects existing in the sparse degree estimation method, a faster and more reasonable estimation method is proposed. And on the basis of analyzing the advantages and disadvantages of several existing sparse reconstruction algorithm, two new algorithms is proposed.Firstly, based on the study of the basic theory of compressed sensing, this paper analyses the signal processing flow using compressed sensing theory, including signal sparseness, the signal sampling and the existing reconstruction algorithms of signal, then analyses the advantages and disadvantages of various reconstruction algorithms with the MATLAB simulation demonstration.Secondly, this paper studies the basic principle of synthetic aperture radar imaging, and discusses some traditional imaging algorithms, such as Range - Doppler algorithm, Chirp Scaling algorithm,Range Migration method,etc. And the simulation results are presented.Then the two commonly used high resolution spectrum estimation algorithms are given and proved. The problems that SAR technology faced, including high requirements of A/D conversion technology and the hardware circuit because of high speed sampling, and the defective radar data can limit the resolution of radar imaging, etc., are pointed out.Finally, this paper introduces the SAR imaging technology based on the compressed sensing,and analysis the current situation of compressed sensing's application in SAR,and introduces the complementary space matching pursuit algorithm in detail. The paper improves the existing orthogonal matching pursuit algorithm and the gradient pursuit algorithm, and then combines them with complementary space matching pursuit algorithm respectively to get two kinds of new signal reconstruction algorithm. The MATLAB simulation results verifies the accuracy of these two kinds of improved algorithms, and they both can achieve high resolution radar images and obvious improved reconstruction speed.In addition, the two new algorithms can compress the SAR images of scene and reconstruct it with better quality, and reduce the pressure of storage and transmission consequently.Through analyzing theories and the simulations, we proves that applying the compressed sensing theories to radar imaging system can solve the problems of high speed sampling and obtain the images with high resolution. Therefore the SAR imaging algorithm based on compressed sensing has practicability in radar imaging systems.