摘要
针对飞机目标的分类问题,提出了一种双辨别子空间高分辨距离像雷达目标识别方法。该方法首先依据Fisher准则导出距离像总散布矩阵的零空间中不含有辨别信息的结论,利用这一结论,对类间和类内散布矩阵进行预降维,降低了后续计算的复杂度。从全局的角度出发,基于类内散布矩阵零空间与非零空间所包含的辨别信息分别建立辨别子空间,实现对目标的特征提取。对ISAR实测飞机数据进行了分类,并与经典子空间方法进行比较,结果表明所提算法有效改善了目标识别性能。
A double discriminant subspaces (DDS)algorithm for high resolution range profile radar target recognition is proposed. Firstly, one conclusion that there exists no useful discriminative information in the null space of the population scatter matrix is derived through the Fisher's criterion, which can be used to reduce the di- mensionality of the original between-and within-class scatter matrix as well as the computation complexity of the following operation. Then from global viewpoint,it carries out feature extraction by making full use of the discriminative information in both null space and non-null space of the within-class scatter matrix. Experiments on three measured airplanes data together with a comparison to several classical subspace methods are conducted,and the results show that DDS algorithm effectively improves the classification performance.
出处
《电波科学学报》
EI
CSCD
北大核心
2009年第5期934-938,共5页
Chinese Journal of Radio Science
基金
国家自然科学基金资助项目(No.60702070)
关键词
雷达目标识别
高分辨距离像
双辨别子空间
特征提取
radar target recognition
high resolution range profile(HRRP)
double discriminant subspaces
feature extraction