针对雷达目标识别问题，提出了一种完全鉴别分析特征提取方法。首先依据Fisher准则导出样本总散度矩阵的零空间不含有鉴别信息的结论，利用这一结论，对类间和类内散度矩阵进行预降维，降低了后续计算的复杂度。然后基于类内散度矩阵零空间与非零空间所包含的鉴别信息分别建立子空间，实现对目标的特征提取。对三类飞机目标实测回波数据的识别结果表明了所提方法的有效性。%A Complete Discriminant Analysis (CDA) 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 total scatter matrix is derived from the Fisher's criterion, which can be used to reduce the dimensionality of scatter matrices, thus improve the computation efficiency of the following computation. Then it carries out feature extraction based on two discriminant subspaces, which makes full use of the discriminative information in both null-and non-null space of the within-class scatter matrix respectively. Experiments on the measured data are conducted, and the results confirm the effectiveness of the proposed method.
展开▼
