多重测量矢量模型下的稀疏步进频率SAR成像算法

发布时间：2019-01-26 14:06

【摘要】：基于压缩感知(Compressed Sensing,CS)的合成孔径雷达(SAR)成像算法可以用低于Nyquist采样率的采样数据完成稀疏目标高分辨成像。然而已有的算法在重构1维距离像时采用的大都是单重测量矢量(Single Measurement Vectors,SMV)模型,存在着重构耗时长、受噪声干扰大的缺点。该文从压缩感知的多重测量矢量(Multiple Measurement Vectors,MMV)模型出发,利用多重测量矢量恢复具有相同稀疏结构的联合稀疏目标信号源,从理论与实验角度分析了基于MMV模型的SAR 1维距离像成像性能,提出了一种距离向基于MMV模型,方位向基于SMV模型的2维SAR成像算法。该算法从耗时上、重构精度上均优于SMV模型下的CS成像算法。通过对仿真数据和地基雷达实测数据的处理,验证了算法的有效性。
[Abstract]:The synthetic Aperture Radar (SAR) (SAR) imaging algorithm based on compressed sensing (Compressed Sensing,CS) can use the sampling data below the Nyquist sampling rate to complete the high resolution imaging of sparse targets. However, most of the existing algorithms used in the reconstruction of 1-D range profile are single measurement vector (Single Measurement Vectors,SMV) models, which have the disadvantages of long reconstruction time and large noise interference. Based on the compressed sensing multiplex measurement vector (Multiple Measurement Vectors,MMV) model, the joint sparse target signal source with the same sparse structure is recovered by using the multiple measurement vectors. In this paper, the imaging performance of SAR 1-D range profile based on MMV model is analyzed theoretically and experimentally, and a 2-D SAR imaging algorithm based on MMV model and SMV model in azimuth direction is proposed. In terms of time consuming and reconstruction accuracy, the algorithm is superior to the CS imaging algorithm based on SMV model. The validity of the algorithm is verified by processing the simulation data and the measured data of ground-based radar.
【作者单位】：空军工程大学信息与导航学院;复旦大学电磁波信息科学教育部重点实验室;
【基金】：国家973计划项目(2010CB731905) 国家自然科学基金(61172169)资助课题
【分类号】：TN957.52

【参考文献】