基于聚类分析和神经网络的雷达目标三维散射中心提取  

3-D Scattering Center Extraction Based on Clustering Analysis and Neural Network

作　　者：吕玉增[1] 任双桥[1] 刘永祥[1] 曹敏[1] 黎湘[1] 

机构地区：[1]国防科技大学ATR国家重点实验室,湖南长沙410073

出　　处：《信号处理》2006年第4期462-466,共5页Journal of Signal Processing

基　　金：国防预研项目(No.4130304-01)资助课题

摘　　要：本文根据目标散射中心在雷达视线上的几何投影关系,分析了由多姿态角下一维散射中心重构雷达目标三维散射中心的基本原理和散射中心关联等技术难点,提出了基于聚类分析和神经网络两种三维散射中心提取方法并比较了二者特点。最后,本文采用GTD散射模型生成的回波数据和暗室测量数据进行实验,验证了两种方法具有实现简单、结果稳定等优点。Based on the geometrical projection of scattering center on the line of radar sight, a technique for reconstructing the 3- D scattering centers of a target is proposed in the paper, using muhiplel-D scattering centers in different aspect angles. After analyzing the principle of reconstructing and the mechanism for scattering centers association, two methods based on clustering analysis and neural network are provided and then compared. The experiments based on simulation and measurement data are developed, and experimental result shows the validity and robustness of performed algorithms.

关 键 词：聚类分析 神经网络 关联 三维散射中心 

分 类 号：TN957.52[电子电信—信号与信息处理] TN957.51[电子电信—信息与通信工程]