Research On The Extracting Methods Of The Scattering Center Feature From Wideband Radar Echo

With the development of theoretical research and the improvement of the resolution of the radar signal, wideband High-resolution radar target recognition has become an important development direction of radar target recognition. The target echo of wideband high-resolution radar is no longer seen as "point", but a high-dimensional signal composed by the target scatters distribution at different distances. With the research about these scattering points, not only can get the position of target, but also can derive the essential attributes to reflect the objective characteristics such as size and structure. Based on the wideband high-resolution radar echo data, the target feature extraction methods has been studied. The main work can be summarized as follows:1. Four sparse algorithms have been researched, and the alternating projection method is introduced to optimization algorithm. The basic principles and operating procedures of four sparse algorithms have been studied. Combined with experiment give their apply results. A method combine with alternating projection has been proposed. This method can improve the sparse solver and robustness to noise.2. The one-dimensional simple scattering center extraction methods were studied. The Geometrical Theory of Diffraction model of the scattering center was introduced, and the basic concept of polarization scattering matrix was described. A method based on high resolution range profile extract multi-polarization scattering centers was proposed. Experimental results show that this method can accurately extract the target scattering center position and polarization scattering matrix.3. The Attributed scattering center extraction methods were discussed.(1) A method of attributed scattering centers extraction combined alternating projection and orthogonal matching pursuit was presented. By using the target supporting SAR image region this method can achieving the goal of automatic initialization parameter, reducing the dimensions of the dictionary and reduces the computational complexity. Because of the alternating projection method, this method can reduces the interference caused by the coupling between the target attribute scattering centers, improves the accuracy of the algorithm to obtain the target features.(2) A method of frequency domain of attributed scattering centers extraction combined image information was presented. This method uses the image information of target to achieve the goal of decoupling between the scattering centers, can reduces the computational complexity and improve the accuracy of the extracted parameters.