当前位置: X-MOL 学术IEEE Trans. Geosci. Remote Sens. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
MCANet: A Multidimensional Channel Attention Residual Neural Network for Pansharpening
IEEE Transactions on Geoscience and Remote Sensing ( IF 7.5 ) Pub Date : 9-12-2022 , DOI: 10.1109/tgrs.2022.3205626
Dajiang Lei 1 , Peng Chen 1 , Liping Zhang 1 , Weisheng Li 1
Affiliation  

In the remote sensing image fusion field, fusion methods based on deep learning (DL) are the latest techniques in panchromatic sharpening (pansharpening). However, existing pansharpening methods based on neural networks cannot adequately inject the spatial feature information of panchromatic (PAN) images into fusion images, and they do not exploit the feature relationships between spatial locations, such as rows and columns of feature maps. To solve these problems, a multidimensional channel attention residual neural network (MCANet) is proposed in this article. To preserve the structural information in PAN images, a two-stream detail injection (TSDI) module is proposed, and the local skip connection operation is adopted to mine more spectral and structural information. A multidimensional channel attention (MCA) module is also designed to enable the network to learn the nonlinear mapping relationships between image spatial locations. In addition, a multiscale feature fusion (MSFF) module is designed to improve feature representation in the image fusion process, which is conducive to improving the pansharpening effect. The experimental results on the WorldView-2 (WV-2), GaoFen-2 (GF-2), and QuickBird (QB) datasets demonstrate that the proposed method outperforms state-of-the-art methods both visually and quantitatively.

中文翻译:


MCANet:用于全色锐化的多维通道注意残差神经网络



在遥感图像融合领域,基于深度学习(DL)的融合方法是全色锐化(pansharpening)的最新技术。然而,现有的基于神经网络的全色锐化方法不能充分地将全色(PAN)图像的空间特征信息注入到融合图像中,并且它们没有利用空间位置之间的特征关系,例如特征图的行和列。为了解决这些问题,本文提出了多维通道注意力残差神经网络(MCANet)。为了保留PAN图像中的结构信息,提出了双流细​​节注入(TSDI)模块,并采用局部跳跃连接操作来挖掘更多的光谱和结构信息。还设计了多维通道注意(MCA)模块,使网络能够学习图像空间位置之间的非线性映射关系。此外,还设计了多尺度特征融合(MSFF)模块来改善图像融合过程中的特征表示,有利于提高全色锐化效果。 WorldView-2 (WV-2)、GaoFen-2 (GF-2) 和 QuickBird (QB) 数据集上的实验结果表明,所提出的方法在视觉和定量方面均优于最先进的方法。
更新日期:2024-08-28
down
wechat
bug