Technology Research, Based On Synthetic Aperture Radar Target Recognition

SAR (Synthetic Aperture Radar) ATR (Automatic Target Recognition) is crucial to the success of battlefield awareness and has become a very hot research topic. In recent years, radar target recognition has made steady progress in many fields, including feature extraction, target classification and recognition. Some ATR systems have been built and have been successfully used in the areas like ground detecting and precision guidance in spaceborne/airborne SAR.This thesis first reviews the ATR fundamentals and the state-of-the-art development of SAR ATR techniques.Then the paper covers four parts: power transform, SAR image enhancement, SAR pose estimation and SAR target recognition. Main contributions include:Firstly, Research on the normality of power transform in radar target recognition is commented. The influence of power transform on the data of Gamma,Rayleigh and other distributions is discussed .The normality of experiment results are tested by the methods of Kurtosis and Skewness test and the Pearsons chi-square test. The test result proves that the power transform is effective.Secondly, SAR image enhancement based on image regularization of potential functions is proposed. In this method, the target function of SAR image is designed, and then SAR image enhancement is converted to the optimization of target function. Potential function is used as the cost function of regularization, and then identity matrix is used to take place the projection matrix of SAR image model. Many experimental results of MSTAR public target database show that this method is feasible and it can obtain the approving result.Thirdly, Radon transform and 2-D continuous wavelet transform are applied to recognize the target pose.The angular energy density of 2-D continuous wavelet transform combined with Radon Transform is proposed. Many experimental results of MSTAR public target database show that this method is feasible and it can obtain the pose correctly.Finally, according to the characteristic of SAR target image, several SAR auto target recognition methods are proposed. They are Guass model, gamma model and log model. The distribution of SAR image pixels are made use of to gain the parameters of the model which can confirm the SAR recognition database. Regularization enhancement and pose estimation are used to decrease the calculation of the algorithm efficiently. Many experimental results of MSTAR public target database show that this method is feasible and it can obtain the approving result.