基于高分辨一维像的目标特征提取方法  

Extraction of Target Feature Using High Resolution Range Profile

作　　者：徐庆[1] 王秀春[1] 李青[1] 

机构地区：[1]南京电子技术研究所,南京210013

出　　处：《现代雷达》2009年第6期60-63,共4页Modern Radar

摘　　要：雷达高分辨一维像反映了目标的精细结构和材料质地,但它具有姿态敏感性和平移敏感性,直接利用一维像进行目标识别难以取得理想的效果。文中基于雷达高分辨一维像,提出了能量聚集区长度、强散射中心数目、一维像序列的方差和"信源熵"等4个具有平移不变性的特征概念,将其构成特征向量,并用实测数据验证其可行性和有效性,取得了较好的效果。High resolution range profile （HRRP） shows the fine structure and texture of the target, but it is sensitive to the aspect and the translation of target generally, it is difficult to obtain a satisfactory result by applying the HRRP to target recognition directly. In this paper, four features based on HRRP are proposed and formed into a feature vector ,which are length of energy concentrated area, the number of strong scattering centers,variance and entropy of HRRP. Finally, the real data are used to show the feasibility and for validation when using the feature vector proposed in this paper and good result is obtained.

关 键 词：高分辨一维像 姿态敏感性 平移敏感性 特征提取 

分 类 号：TN957.52[电子电信—信号与信息处理]