提出了一种基于最优聚类中心的雷达目标一维距离像识别方法。该方法利用训练数据集建立最小平方距离准则下的最优变换矩阵 ,使用该变换矩阵可增大同类目标的特征聚合性 ,从而减少同类之间差异 ,同时 ,通过在子像空间选定一组最优聚类中心来增大异类目标特征的可分离性 ,加大异类之间差异 ,提高雷达目标识别率。仿真实验结果表明了该方法的有效性。 A method of radar target one-dimensional distance image recognition based on optimal cluster center is proposed. This method uses the training dataset to establish the optimal transformation matrix under the minimum square distance criterion. The transformation matrix can be used to increase the characteristic aggregation of similar targets and reduce the differences among similar objects. At the same time, Optimal clustering center to increase the separability of heterogeneous target features, increase the differences between the heterogeneous and improve the radar target recognition rate. Simulation results show that the method is effective.