基于高分辨距离像的雷达目标识别的关键是如何提取距离像的有效鉴别特征。通过分析核Fisher准则函数下核类内散布矩阵的零空间信息与非零空间信息对识别的贡献,基于同构映射思想,求解出一组最优鉴别矢量集,用于提取距离像的鉴别特征。同时,给出一种快速算法,用较少的训练样本表达投影方向,加快了鉴别特征的提取速度。基于外场实测数据的实验证明了方法的有效性。 The key point of radar target recognition based on high resolution range image is how to extract the effective distinguishing features of distance image. By analyzing the contributions of the zero space information and the non-zero spatial information of the scatter matrix within the kernel Fisher criterion function to the recognition, based on the idea of ​​isomorphic mapping, a set of optimal discriminant vector sets are obtained to extract the discriminant features of distance images . At the same time, a fast algorithm is proposed, which can express the projection direction with fewer training samples and speed up the extraction speed of the identification features. Experiments based on measured data show that the method is effective.