现代防御技术2013，Vol.41Issue(4)：126-130,5.DOI:10.3969/j.issn.1009-086x.2013.04.023
一种SAR图像特征提取和目标分类的新方法
New Method for Synthetic Aperture Radar Images Feature Extraction and Target Classification
摘要
Abstract
A new method for synthetic aperture radar images feature extraction and target recognition based on Kernel PCA in wavelet domain and support vector machine is presented.After two-dimension wavelet decomposition of a SAR image,feature extraction is implemented by picking up Kernel principal component of the low-frequency sub-band image.Then,support vector machine is used to perform target recognition.Using MSTAR SAR data to experiment,results show that correctness of recognition is enhanced obviously.关键词
合成孔径雷达/二维离散小波变换/核主成分分析/支持向量机/自动目标识别Key words
synthetic aperture radar (SAR)/ two-dimension wavelet transform/ kernel principle component analysis PCA/ support vector machine(SVM) / automatic target recognition(ATR)分类
信息技术与安全科学引用本文复制引用
李勇,王德功,常硕,关春健..一种SAR图像特征提取和目标分类的新方法[J].现代防御技术,2013,41(4):126-130,5.
