基于支持向量机的浅地层探地雷达目标分类识别研究

对于探地雷达用于探测地雷情况下的目标识别,从实用性上说只需识别地雷或非地雷两类目标即可,而通常的支持向量机正是用于两类分类.本文结合浅地层探地雷达数据的特点,提出了一种基于支持向量机的浅地层探地雷达目标分类识别方法,并分析了时域,傅立叶谱,及离散小波变换三种特征数据用于所提方法时的效果.通过对实测数据进行处理,结果表明三种特征数据用于所提方法都能取得较好的效果.

Research on Ground Penetrating Radar Target Identification Based on Support Vector Machines in Shallow Subsurface Application

Ground penetrating radar(GPR) is an effective tool to detect mine.Ground penetr a ting radar target identification needs to distinguish mine and non-mine only,a nd support vector machines(SVM) can be used to two-class classification pr oblem in practice.This paper analyzed the characteristic of shallow subsurface G PR data,proposed a method of GPR target identification based on support vector m achines,and processed measurement data by using the proposed method in time sam p le,Fourier spectra,and discrete wavelet transform as feature data.The results of processing me asurement data show this method is effective.