基于改进多重测量向量模型的SAR成像算法  

A Novel SAR Imaging Algorithm Based on Modified Multiple Measurement Vectors Model

作　　者：陈一畅 张群[1,2] 杨婷[3,4] 罗迎[1] 

机构地区：[1]空军工程大学信息与导航学院 [2]复旦大学电磁波信息科学教育部重点实验室 [3]空军预警学院黄陂士官学校 [4]中国人民解放军95133部队

出　　处：《电子与信息学报》2016年第10期2423-2429,共7页Journal of Electronics & Information Technology

基　　金：国家自然科学基金(61471386);中国博士后基金(2015M570815);陕西省统筹创新工程-特色产业创新链项目(2015KTTSGY04-06)~~

摘　　要：近年来,基于压缩感知(Compressed Sensing,CS)理论的稀疏场景SAR成像成为研究热点。在CS理论中,对于具有相同稀疏结构的联合稀疏目标信号源,多重测量向量(Multiple Measurement Vectors,MMV)模型可以获得优于单重测量矢量(Single Measurement Vector,SMV)模型的重构性能。然而,在距离徙动(Range Migration)不可忽略的应用场景,SAR各脉冲回波1维距离像具有不完全相同的稀疏结构,这使得无法在SAR应用中直接引入MMV模型。该文针对MMV模型与SAR距离徙动的矛盾,提出一种改进的MMV模型。在该模型下,各方位向位置对应的1维距离像的稀疏结构不要求完全相同,而是符合距离徙动特性。论文所提出的RM-OMP算法根据符合距离徙动特性的稀疏结构搜索稀疏信号支撑集,可以准确地重建稀疏信号源。论文采用仿真数据和实测数据实验验证了所提模型和算法的有效性。Recently, the Compressed Sensing （CS） theory becomes the researching hot point in SAR imaging. The Multiple Measurement Vectors （MMV） model of CS theory can be used to effectively represent the jointly sparse signals, and it can obtain better performance than Single Measurement Vector （SMV） model. Because the SAR range profiles at different pulses have different sparse structures, which result in the MMV model can not be directly used in the scenario of synthetic aperture radar imaging. In this paper, a modified MMV model is proposed for SAR imaging, and the Range Migration （RM） effect is embedded into the proposed model. Correspondingly, a modified Orthogonal Matching Pursuit （OMP） algorithm is developed to obtain the high-resolution range profile. Experiments based on simulated and measured data demonstrate the validity of the proposed model and the algorithm.

关 键 词：合成孔径雷达 稀疏恢复 多重测量向量模型 距离徙动 

分 类 号：TN957.52[电子电信—信号与信息处理]