重庆理工大学学报（自然科学版）2018，Vol.32Issue(5)：204-209,6.DOI:10.3969/j.issn.1674-8425(z).2018.05.031
卷积神经网络在SAR目标识别中的应用
Application of Convolution Neural Network in SAR Target Recognition
摘要
Abstract
Aiming at the problem of SAR target recognition,this paper proposes a SAR target recognition method based on convolution neural network,and improves the algorithm on this basis.It proposes a new model called CNN-SVM,and the soft-max classifier of traditional convolution neural network is replaced by support vector machine,and it classifies the features extracted by convolution neural network.Firstly,it cuts and de-noises the samples,and then it expands the samples by adding noise and denoising etc.Simulation experiments on MSTAR datasets show the correct recognition rate of traditional convolution neural network and the improved convolution neural network is 97.5% and 99.4%,respectively,which proves the effectiveness of the proposed algorithm.关键词
SAR目标识别/卷积神经网络/支持向量机Key words
SAR target recognition/convolution neural network/support vector machine分类
信息技术与安全科学引用本文复制引用
郝岩,白艳萍,张校非,杜敦伟..卷积神经网络在SAR目标识别中的应用[J].重庆理工大学学报（自然科学版）,2018,32(5):204-209,6.基金项目
国家自然科学基金资助项目(61774137) （61774137）
山西省自然科学基金资助项目(201701D22111439,201701D221121) （201701D22111439,201701D221121）
山西省回国留学人员科研项目(2016-088) （2016-088）
