基于核支持向量最优变换矩阵的雷达目标一维距离像识别

提出了一种基于支持向量最优变换矩阵的雷达目标一维距离像识别方法。该方法利用支持向量构建类间散布矩阵和类内散布矩阵,结合零空间特性得到最优变换矩阵,该变换矩阵被用来从原始一维距离像中提取判别特征。对输入目标,利用欧式距离进行分类,以确定目标所属类别。对三类飞机的实测一维距离像数据进行了仿真实验,实验结果表明了该方法的有效性。

Radar Target Recognition Using Range Profiles Based on KSVs Optimal Transform Matrix

The paper proposes a novel approach for radar target recognition based on Kernel Support Vectors(KSVs) optimal transform matrix,which constructs a between-class matrix and a within-class scatter matrix by use of KSVs.In addition,the null-space fisher method is exploited to calculate the optimal transform matrix,which is used to extract the discriminant features form the original range profiles.For the test sample,final decision is made in accord with the Euclidean distance.Experimental results on range profiles of three kinds of planes demonstrate the effectiveness of this proposed method.