基于小波分解及KCN的雷达目标特征提取  

Radar target feature extraction using wavelet transform and KCN

作　　者：王晓丹[1] 王积勤[1] 

机构地区：[1]空军工程大学导弹学院,陕西三原713800

出　　处：《电波科学学报》2003年第1期32-37,共6页Chinese Journal of Radio Science

基　　金：空装国防预研基金课题资助 (40 2 0 70 2 0 1)

摘　　要：讨论了利用一维距离像进行雷达目标识别的特征提取。从缩减表征距离像的特征维数和表征目标所需特征数出发 ,提出了一种基于小波分解及KCN的目标特征提取方法。为比较不同特征提取方法及所提取特征的性能 ,引入了Fisher距离来表明各类的类内紧致性和类间分离性。实验结果表明能够有效地进行特征提取。Feature extraction of radar target recognition using high resolution range profiles is discussed in this paper. A feature extraction scheme based on wavelet transform and Kohonen clustering network(KCN) is proposed from the points of reducing feature dimension and feature numbers to representing a target. To compare the performance of different feature extraction schems, the Fisher distance measure is introduced to reveal the in-class compactness and between-class separateness. Experiment result show that the presented scheme can extract the target feature effectively.

关 键 词：小波分解 雷达 目标识别 特征提取 KCN Fisher距离 类间分离性 

分 类 号：TN95[电子电信—信号与信息处理]