采用DBSCAN改进的矩阵束极点提取算法  

A Matrix Pencil Method Algorithm Improved by DBSCAN to Poles Extraction

作　　者：邢煦然 赵宏钟 贾鑫 XING Xuran 1, ZHAO Hongzhong 2, JIA Xin 2(1.Department of Graduate Management, Equipment Academy, Beijing 101416, China;2.Department of Optical and Electronic Equipment, Equipment Academy, Beijing 101416, China)

机构地区：[1]装备学院研究生院,北京101416 [2]装备学院光电装备系,北京101416

出　　处：《雷达科学与技术》2018年第3期327-332,共6页Radar Science and Technology

摘　　要：极点是谐振区雷达目标的最主要特性。矩阵束法是目前提取雷达目标极点中应用最广泛的算法。在使用矩阵束法提取极点时,现有阶数估计方法普遍存在阶数偏大和虚假极点难以区分的问题。针对这一问题,首先描述了矩阵束法估计得到的极点的分布特性;然后通过对极点集投影估计了真实极点的个数;最后用DBSCAN算法对极点分类,剔除了虚假极点,保留了真实的极点。通过FEKO仿真细杆和球体目标的电磁散射数据验证了算法,有效剔除了虚假极点并提取出了真实极点。Poles are the main feature of the radar target in the resonance region.The matrix pencil method（MPM）is the most widely used algorithm in poles extraction.When MPM is used to extract poles of radar target,the estimated order is usually bigger than the number of poles,which leads to false poles mixture with real poles.In this paper,the distribution character of poles extracted by MPM is discussed.Then,a method of estimating the poles number by the projection of pole set is put forward.At last,DBSCAN is used to classify the poles,the false poles are eliminated and the real poles are reserved.Data of two targets created by FEKO are used to prove the effectiveness of the proposed algorithm.The algorithm can remove the false poles effectively.

关 键 词：密度聚类 极点 矩阵束法 谐振区 

分 类 号：TN957[电子电信—信号与信息处理]