雷达目标一维散射中心识别特征提取研究

Research on one-dimensional scattering center feature extraction from radar targets

针对雷达目标高分辨一维距离像在目标识别中存在计算量和存储量过大的问题,研究了基于一维散射中心的目标识别特征。在使用矩阵束法提取散射中心的基础上,定义了散射中心阶数、散射中心相对位置和中心矩三个平移不变特征用于目标识别,较好地描述了目标散射特性。三类目标外场实测数据的识别试验结果显示,相对于单一特征的目标识别,综合上述三个特征的识别不但能获得更好的目标识别率,而且识别结果对高斯白噪声不敏感。

The extraction of target features is investigated for radar high resolution range profile （HRRP） recognition. According to the scattering centers obtained with the matrix-pencil algorithm, three translation-invariant features of the order of scattering center, the relative position of scattering center and central moments are designed to classify the airplane objects. The experiment comparisons based on the measured data show that proposed method combined with these three features achieves good classification performance and is low sensitive to Gausses white noise.