电子测量与仪器学报

杜晨光,胡建文,胡佩.半监督卷积神经网络遥感图像融合[J].电子测量与仪器学报,2021,35(6):63-70

半监督卷积神经网络遥感图像融合

Semi-supervised convolutional neural network remote sensing image fusion

DOI：

英文关键词:Convolutional neural network semi-supervision remote sensing image fusion spectral degradation network spatial degradation network

基金项目:国家自然科学基金项目（61601061）、湖南省教育厅项目（14B006）、电力机器人湖南省重点实验室开放研究课题（PROF1902）资助项目

作者	单位
杜晨光	1.长沙理工大学电气与信息工程学院
胡建文	1.长沙理工大学电气与信息工程学院
胡佩	1.长沙理工大学电气与信息工程学院

Author	Institution
Du Chenguang	1.School of Electrical and Information Engineering, Changsha University of Science and Technology
Hu Jianwen	1.School of Electrical and Information Engineering, Changsha University of Science and Technology
Hu Pei	1.School of Electrical and Information Engineering, Changsha University of Science and Technology

摘要点击次数: 433

全文下载次数: 4

中文摘要:

近几年随着深度学习的发展,学者们利用卷积神经网络实现遥感图像融合取得了不错的效果。由于没有高分辨率多光谱图像作为参考图像,所以一般基于深度学习的方法在退化图像上训练模型,然后用训练好的模型去预测高分辨率多光谱图像,但是退化图像的融合过程并不能完全反映原始图像的融合过程。为了改善融合性能,提出了一种半监督卷积神经网络遥感图像融合方法,该方法在退化图像和原始图像上使用同一个融合网络同时进行训练。退化图像的融合具有相应的参考图像,采用常规的监督学习方式对融合网络进行训练,还加入了光谱损失来更好的保持光谱信息。而原始图像的融合不存在高分辨率多光谱参考图像,设计了光谱退化网络和空间退化网络对融合图像进行退化,再训练融合网络。实验结果表明,提出的方法光谱与细节保真效果好,优于对比方法。

英文摘要:

With the development of deep learning in recent years, remote sensing image fusion methods based on convolutional neural network were proposed and presented with good performance. Because there is no high-resolution multispectral image as a reference, the convolutional neural network is trained in the degraded images. The trained network is used to predict high resolution multispectral images. However, the fusion process of degraded images cannot reflect the fusion process of original images. In order to improve fusion performance, a semi-supervised fusion method based on convolutional neural network is proposed. The same fusion network is trained in the degraded image and the original image simultaneously. Because degraded image fusion has the corresponding reference image, the supervised learning method is used to train the fusion network. Moreover, the spectral loss is added to preserve the spectral information. However, there is no high-resolution multispectral reference image in the original image fusion. Spectral degradation network and spatial degradation network are designed to train the fusion network. The experimental results show that the proposed method is better than the compared method in preserving the spectral and details.

查看全文查看/发表评论下载PDF阅读器