基于高斯字典原子稀疏表示的高精度宽角SAR成像方法  

High-precision wide angle SAR imaging method based on sparse representation of Gaussian dictionary atoms

作　　者：陈晨[1,2,3] 魏中浩 徐志林[1,2,3] 张冰尘 CHEN Chen;WEI Zhonghao;XU Zhilin;ZHANG Bingchen(Institute of Electronics,Chinese Academy of Sciences,Beijing 100190,China;Key Laboratory of Technology in Geospatial Information Processing and Application System,Beijing 100190,China;University of Chinese Academy of Sciences,Beijing 100190,China)

机构地区：[1]中国科学院电子学研究所,北京100190 [2]空间信息处理与应用系统技术重点实验室,北京100190 [3]中国科学院大学,北京100190

出　　处：《系统工程与电子技术》2019年第11期2471-2478,共8页Systems Engineering and Electronics

基　　金：国家自然科学基金（61571419）资助课题

摘　　要：为了提高从宽角合成孔径雷达(synthetic aperture radar,SAR)图像中提取目标后向散射各向异性特性的性能,在宽角SAR字典稀疏表示模型的基础上,提出一种基于高斯字典原子的高精度宽角SAR成像方法。在字典构造上,采用不同中心位置、相同方差的高斯函数。在求解稀疏表示系数上,采用广义最小最大凹惩罚稀疏重构算法求解。最后,根据稀疏表示系数的重构结果以及构造的字典得到目标的后向散射各向异性特性。通过仿真实验和Backhoe数据对算法进行验证,结果表明,该方法能够高精度地提取目标的后向散射各向异性特性。In order to improve the performance of extracting the anisotropic properties of target backscattering from wide angle synthetic aperture radar images,a high-precision wide angle synthetic aperture radar imaging method based on sparse representation of Gaussian dictionary atoms is proposed.In the dictionary construction,the method uses Gaussian functions with different center positions and the same variance.In solving the sparse representation coefficients,the method uses the generalized minimax concave penalty sparse reconstruction algorithm.Finally,the anisotropic properties of the target backscattering are obtained according to the reconstruction result of the sparse representation coefficients and the constructed dictionary.The method is validated by simulation experiments and the Backhoe data.The results show that the proposed method can extract the anisotropic properties of the target backscattering with high precision.

关 键 词：宽角合成孔径雷达 高斯字典原子 稀疏表示 广义最小最大凹惩罚 

分 类 号：TN95[电子电信—信号与信息处理]