基于黎曼流形监督降维的矩阵CFAR增强检测  

Enhanced Matrix CFAR Detection Based on Supervised Dimensionality Reduction of Riemannian Manifold

作　　者：杨政 程永强 吴昊[1] 黎湘[1] 王宏强[1] YANG Zheng;CHENG Yongqiang;WU Hao;LI Xiang;WANG Hongqiang(College of Electronic Science,National University of Defense Technology,Changsha,Hunan 410073,China)

机构地区：[1]国防科技大学电子科学学院,湖南长沙410073

出　　处：《信号处理》2021年第11期2013-2021,共9页Journal of Signal Processing

基　　金：国家自然科学基金(61871472)。

摘　　要：矩阵CFAR检测是从几何流形角度处理雷达目标检测问题的新技术。为进一步提升其在复杂杂波背景下的检测性能,本文提出一种黎曼流形监督降维的矩阵CFAR增强检测方法。首先,将检测问题视为目标与杂波的分类问题,分别构建黎曼流形上目标单元与杂波单元的类内和类间权重矩阵;其次,为增强目标与杂波的可分性,采用保持类内几何距离最小,类间几何距离最大的准则建立降维目标函数,并基于Grassmann流形求解降维优化问题获得映射矩阵;最后,提出一种矩阵CFAR增强检测方法,实现目标增强检测。采用蒙特卡罗方法对仿真数据和实测海杂波数据进行实验分析,结果表明,所提出的方法能够进一步提升检测性能。Radar target detection problem can be processed by a new thinking and method called matrix CFAR detection based on geometry manifold.An enhanced matrix CFAR detection method based on supervised dimensionality reduction of Riemannian manifold is proposed to further improve its performance in complex clutter background in this paper.Firstly,the detection problem is regarded as a classification problem of the target and clutter,and within-class weight matrix and between-class weight matrix of the target and clutter on the Riemannian manifold are constructed,respectively.Then,in order to enhance the discriminability between the clutter and target,an optimal objective function for dimensionality reduction is obtained by maintaining the minimum within-class geometry distance and the maximum between-class geometry distance,and an explicit mapping matrix can be obtained by solving the optimization problem on Grassmann manifold.Finally,an enhanced matrix CFAR detection method is proposed for the enhancement detection of the target.By Monte Carlo technique,the experimental results of simulated data and real sea clutter data demonstrate that the proposed method can achieve a better detection performance.

关 键 词：矩阵CFAR增强检测 黎曼流形 监督降维 GRASSMANN流形 映射矩阵 

分 类 号：TN957.51[电子电信—信号与信息处理]