针对雷达高分辨距离像(HRRP)的方位敏感性和平移敏感性,对一定角域内的HRRP非相干平均,提取具有平移不变性的中心矩作为特征向量,采用Karhunen2Loeve变换进一步进行特征压缩,建立相应的支撑矢量机(SVM)分类算法。与基于原始距离像特征的最大似然(ML)方法和基于中心矩特征的ML方法识别结果比较,该方法在减少计算量的同时具有较高的识别率,具有良好的推广能力。 According to the azimuth sensitivity and translational sensitivity of HRRP, a non-coherent HRRP HRRP is averaged. The center moment with translational invariance is extracted as the eigenvector. The Karhunen2Loeve transform is used to further perform feature compression. Support Vector Machine (SVM) classification algorithm. Compared with the maximum likelihood (ML) method based on the original distance-like features and the ML method based on the central moment feature, the proposed method has a higher recognition rate while reducing the computational complexity and has good popularization ability.