多级可逆网络在SAR图像识别中的应用  

Application of Multi-level Reversible Network in SAR Image Recognition

作　　者：李青 柯文宇 翟懿奎[1] LI Qing;KE Wen-yu;ZHAI Yi-kui(School of Electronics and Information Engineering,Wuyi University,Jiangmen 529020,China)

机构地区：[1]五邑大学电子与信息工程学院,广东江门529020

出　　处：《五邑大学学报（自然科学版）》2024年第2期40-47,共8页Journal of Wuyi University(Natural Science Edition)

基　　金：广东省基础与应用基础研究基金项目(2021A1515011576);广东省国际科技合作领域专项项目(2021A0505030080、2021A0505060011);广东省教育厅高校计划重点科研项目(2020ZDZX3031、2022ZDZX1032、2023ZDZX1029);江门市基础与应用基础研究重点项目(2220002000246);五邑大学港澳联合研发基金项目(2022WGALH19)。

摘　　要：针对传统卷积神经网络目标识别方法在处理合成孔径雷达图像时存在的信息丢失等问题,我们引入了数据增强技术和多级可逆网络:采用数据增强方法,增加训练样本数量;通过多级可逆网络无损地提取合成孔径雷达图像特征;由交叉熵损失函数和二进制交叉熵损失函数联合训练网络,使用运动和静止目标获取与识别数据集,准确率达到99.75%,与现有的方法相比,准确率较高.To solve the problem of information loss in the processing of SAR images by the traditional convolutional neural networks object recognition method,we introduced the data augmentation technology and Multi-level reversible network:the data augmentation method was used to increase the number of training samples;the features of SAR images were extracted losslessly by a Multi-level reversible network;the network was trained by the cross entropy loss function and the binary cross entropy loss function so that the data set of moving and stationary objects obtained and recognized reached an accuracy rare of 99.75%,which is higher than that achieved by using the existing methods.

关 键 词：数据增强 多级可逆网络 合成孔径雷达 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]