基于判别矢量子空间的雷达目标距离剖面像识别

提出了一种基于判别矢量子空间的雷达目标距离剖面像识别方法。判别矢量子空间一方面在分析意义上是较优的，同时该子空间的维数不受目标类别数限制，从而能够提取更有效的目标分类特征，改善目标正确识别率。仿真实验结果表明：该方法识别率高于特征图像方法和正则子空间法。

Range profile recognition of radar target based on discriminant vector subspace

A approach of radar target recognition is proposed. The discriminable vector subspace is constructed by optimal discriminant analysis on the training dataset and the number of this subspace dimensions does not depend on the number of target classes. Therefore, in the sense of separiablity of classes, the extracted features by this approach are superior to the those by eigen-images and canonical subspace methods. The simulated experimental results on three radar targets show that this method is effective.