Kernel Complete Discriminant Analysis Algorithm for Radar Target Recognition Using HRRPs

Currently, kernel-based methods have drawn much attention from the field of pattern recognition as well as radar target recognition. As we all know, kernel discriminant analysis (KDA) is proved that it is a very effective tool used for dimensionality reduction and feature extraction. However, KDA also suffers from the so-called small sample size problem (SSS) which often exists in high-dimensional pattern recognition data. In order to deal with this problem, a complete KDA called kernel complete discriminant analysis (KCDA) is proposed. The new algorithm views the optimal discriminant vectors as a global transform in the feature space, and which 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. Thus it makes KCDA a more powerful dicriminator. Experiments based on the measured airplanes database are conducted to evaluate the effectiveness of the proposed method, and the results show that it can achieve better classification performance.