Simulation And Research On Statistic Inverse Method With Mirrored Interferometric Aperture Synthetic Radiometer

The aperture synthesis technique uses multiple small aperture antenna synthesis larger diameter antenna, which is convenient for processing and equipping on satellite, and it's performance in remote sensing imaging is more better than real aperture antenna. Since now, there are several countries have put it into practice, and some well-known examples contain the MIRAS equipped on SMOS satellite made by the European Space Agency(ESA), the GeoSTAR made by the National Aeronautics and Space Administration(NASA), etc. Because they are both not able to provide a resolution which meets the needs of people, so they don't have a wide using range of applications. Mirrored interferometric aperture synthesis radiometer(MIASR) is a new interferometric technique proposed recent years, which combines the double antenna interference and reflective surfaces interference. It simplifies the structure and the number of the receiver array based on the traditional synthetic aperture system, and it can get better resolution and sensitivity than conventional synthetic aperture in the same array size.As a new imaging technique, MIASR has many places worth studying about brightness temperature image reconstruction method. In addition to obtaining cosine visibility indirect, then reconstruction image brightness temperature by the inverse cosine transform method, the G matrix model in traditional aperture synthesis radiometer(ASR) is also applicable to MIASR. As with the traditional ASR, MIAS is also facing the ill inverse problems as a result of the presence of errors and other non-ideal factors. By means of regularization, we can weaken the degree of ill and get the approximate solution of brightness temperature under certain priori conditions. But the selection of the model parameters is also a very difficult problem. Bayesian statistical inversion method for conventional ASR established based on the assumption that brightness temperature and visibility are approximate Gaussian distribution. It brings the model parameters into the brightness temperature inversion processing, which is a good solution to this problem, and the performance is the same with the regularization algorithm. This is indeed a new train of thought about brightness temperature inversion.In this thesis, a Bayesian statistical model is applied to reconstruct the brightness temperature of the MIAS. We simulate the statistical characteristics of the antenna cross-correlation output data for the sake of verifying that it approximately satisfies a Gaussian distribution. Then, we give the specific presentation of statistical inversion methods as well as treatment process in MIAS system. Based on these results, we simulate and analyze the one-dimensional and two-dimensional MIAS statistical inversion method, and compare with the traditional inversion algorithm performance. Its performance is quite with regularization inversion method, and it can automatically select the optimal model parameters.