二维全极化ESPRIT超分辨特征提取及性能分析  

Two-Dimensional ESPRIT Super-Resolution Feature Extraction Using Fully Polarized Measurements and Its Performance Analysis

作　　者：邢世其[1] 代大海[1] 王雪松[1] 王涛[1] 

机构地区：[1]国防科技大学电子科学与工程学院,湖南长沙410073

出　　处：《电子学报》2009年第12期2681-2687,共7页Acta Electronica Sinica

基　　金：新世纪优秀人才计划(No.NCET-04-0997);国家自然科学基金(No.60672033);国家自然科学基金重点项目(No.60736006)

摘　　要：提出了二维全极化ESPRIT超分辨特征提取方法,推导了二维全极化参数估计的克拉美罗限,并且在此基础上分析了本文方法的工作性能.同已有超分辨算法相比,二维全极化ESPRIT方法的突出优点在于:(1)通过极化和超分辨的联合处理,实现了对目标相干散射矩阵的精确提取.从而为利用极化信息进行目标识别提供了基础;(2)利用全极化信息,改善了数据的信噪比,提高了参数估计的精度;(3)利用信号子空间的旋转不变性,避免了对多维信号子空间的搜索,提高了算法的实时性.仿真和实测数据均证实了该方法的有效性.值得一提的是,文章首次给出了散射中心位置、散射强度、相干相位等参数估计的克拉美罗限与全极化模型参数的关系,这为超分辨成像算法的性能评价提供了理论依据.In this paper,a polarimetric super processing method based on ESPRIT is proposed,and the CRB of scattering centers parameters estimation,including position,intensity and coherent phase are derived.Experiments reveal that the performance of the proposed method is very close to CRB.Superiorities of our method relies on the following processing techniques：（1） joint processing of super-resolution and full polarization,which enables extraction of coherent scattering matrix of target;（2） utilization of fully polarized backscattered data, which improves SNR and enhances the performance of estimation; （3） utilization of the rotational invafiance property of signal subspace, which greatly reduces computational burden by avoiding multi-dimensional search. Simulation and real data experiments validate its effectiveness. The conclusion is meaningful to polarimetric super-resolution imaging, algorithm evaluation, polarization feature extraction and recognition, and high-resolution polarized radar system design.

关 键 词：超分辨 二维ESPRIT 全极化 克拉美罗限 性能分析 

分 类 号：TN95[电子电信—信号与信息处理]