基于支持向量机的高分辨距离像分类法

High Resolution Range Profile Classification Based on Support Vector Machine

支持向量机 (supportvectormachine ,SVM)是新一代学习机 ,具有良好的泛化性能。高分辨距离像(HRRP)分类是雷达复杂目标分类的重要方法。采用SVM作为分类器 ,研究了飞机目标HRRP分类法。设计了相应的预处理算法 ,并提出了结合VapnikChervonenkis维法和留一 (LOO)交叉验证法的参数选择算法。基于 5种飞机缩比模型的HRRP数据 ,比较了SVM分类法和最大相关分类法的性能 ,研究了噪声、训练用方位角采样数和训练样本集的大小对识别性能的影响。实验结果表明 ,SVM在HRRP分类上具有良好的应用前景。

Support vector machine (SVM) is a new generation learning system with good generalization property High resolution range profile (HRRP) classification is important to radar complex target classification In this paper, we apply SVM to aircraft HRRP classification, propose a preprocessing method and present a new SVM model selection scheme combining leave one out (LOO) cross validation method with Vapnik Chervonenkis dimension method Based on the HRRPs of five types of aircraft, the classification performance of the SVM method is compared with that of the maximum correlation classification method, and the influences of noise, number of sampled training azimuths and size of training set on the performance are researched It is demonstrated by experimental results that SVM is promising to better HRRP classification performances \;