基于多极化多特征提取和SVM的目标识别方法

Radar Target Recognition Approach Based on Multi Polarization Multi Target Feature Extraction and SVM

充分有效地利用全极化雷达目标高分辨一维距离像极化信息是雷达目标特征提取领域一个热点问题。提取全极化雷达目标高分辨一维距离像的极化特征、结合其能量聚集区长度以及中心距特征共同构成特征向量,采用支持向量机进行特征分类,从而提出了一种基于多极化多特征提取和支持向量机的目标识别方法。基于转台数据的实验结果表明,提出的基于多极化多特征提取和支持向量机的目标识别方法有助于获得较高的分类正确率。

Full and effective use of complete polarization radar target for high resolution range profile of radar target polarization information is a hot issue in the field of feature extraction. By extracting polari- zation characteristics of complete polarization radar target high resolution range profile, combining with the length of the energy accumulation area and pitch features to make eigenvectors and by utilizing sup- port vector machine to classify features, a target recognition approach is presented based on multi polari- zation and SVM. Data from the turntable based on the experimental shows that this approach helps a- chieve higher classification accuracy.