支持向量机在车辆目标识别中的应用  

Application of SVM in recognition of vehicles

作　　者：冀振元[1] 李晨雷[1] 唐文彦[1] 

机构地区：[1]哈尔滨工业大学电子工程系,黑龙江哈尔滨150001

出　　处：《系统工程与电子技术》2010年第2期284-286,共3页Systems Engineering and Electronics

基　　金：中国博士后科学基金(20080430906);黑龙江省博士后(LBH-Z07111)资助课题

摘　　要：提出了利用支持向量机对战场侦察雷达目标回波信号进行处理,以实现对卡车、坦克等在地面运动的车辆目标进行分类识别的一种新算法。首先对雷达接收到的目标回波信号作频域分析,从中提取待分类目标信号的特征向量,然后利用所建立的支持向量机模型对目标信号作训练和识别,最后与经典谱分析和神经网络的方法作比较,并采用实际数据验证这种识别方法的有效性。A new method of target recognition is put forward,in which echo of moving targets,such as truck and tank,detected by the scout radar is processed by means of support vector machine（SVM）.Firstly,by analyzing echo of targets received by radar in frequency domain,eigenvector of the signal is extracted,then SVM model is used to train and recognize these targets.Finally,by comparing with recognition method of classic spectrum analysis and neural network,this new method is proved to be efficient in virtue of real data.

关 键 词：支持向量机 特征提取 目标识别 

分 类 号：TN957.52[电子电信—信号与信息处理]