Invertible Attention-Guided Adaptive Convolution and Dual-Domain Transformer for Pansharpening

Pansharpening is the process of fusing a multispectral (MS) image with a panchromatic image to produce a high-resolution MS (HRMS) image. However, existing techniques face challenges in integrating long-range dependencies to correct locally misaligned features, which results in spatial-spectral dist...

Full description

Saved in:
Bibliographic Details
Main Authors: Qun Song, Hangyuan Lu, Chang Xu, Rixian Liu, Weiguo Wan, Wei Tu
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10845120/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849716161487306752
author Qun Song
Hangyuan Lu
Chang Xu
Rixian Liu
Weiguo Wan
Wei Tu
author_facet Qun Song
Hangyuan Lu
Chang Xu
Rixian Liu
Weiguo Wan
Wei Tu
author_sort Qun Song
collection DOAJ
description Pansharpening is the process of fusing a multispectral (MS) image with a panchromatic image to produce a high-resolution MS (HRMS) image. However, existing techniques face challenges in integrating long-range dependencies to correct locally misaligned features, which results in spatial-spectral distortions. Moreover, these methods tend to be computationally expensive. To address these challenges, we propose a novel detail injection algorithm and develop the invertible attention-guided adaptive convolution and dual-domain Transformer (IACDT) network. In IACDT, we designed an invertible attention mechanism embedded with spectral-spatial attention to efficiently and losslessly extract locally spatial-spectral-aware detail information. In addition, we presented a frequency-spatial dual-domain attention mechanism that combines a frequency-enhanced Transformer and a spatial window Transformer for long-range contextual detail feature correction. This architecture effectively integrates local detail features with long-range dependencies, enabling the model to correct both local misalignments and global inconsistencies. The final HRMS image is obtained through a reconstruction block that consists of residual multireceptive field attention. Extensive experiments demonstrate that IACDT achieves superior fusion performance, computational efficiency, and outstanding results in downstream tasks compared to state-of-the-art methods.
format Article
id doaj-art-a3eaaaa1abf8470fbd66a443fda80f1a
institution DOAJ
issn 1939-1404
2151-1535
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
spelling doaj-art-a3eaaaa1abf8470fbd66a443fda80f1a2025-08-20T03:13:07ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352025-01-01185217523110.1109/JSTARS.2025.353135310845120Invertible Attention-Guided Adaptive Convolution and Dual-Domain Transformer for PansharpeningQun Song0Hangyuan Lu1https://orcid.org/0000-0002-2829-8553Chang Xu2https://orcid.org/0009-0008-3253-4839Rixian Liu3https://orcid.org/0000-0002-6217-997XWeiguo Wan4https://orcid.org/0000-0002-3537-979XWei Tu5https://orcid.org/0000-0002-2673-2192College of Information Engineering, Jinhua University of Vocational Technology, Jinhua, ChinaCollege of Information Engineering, Jinhua University of Vocational Technology, Jinhua, ChinaHangzhou Consumer Council, Hangzhou, ChinaCollege of Information Engineering, Jinhua University of Vocational Technology, Jinhua, ChinaSchool of Software and Internet of Things Engineering, Jiangxi University of Finance and Economics, Nanchang, ChinaSchool of Big Data Science, Jiangxi Science and Technology Normal University, Nanchang, ChinaPansharpening is the process of fusing a multispectral (MS) image with a panchromatic image to produce a high-resolution MS (HRMS) image. However, existing techniques face challenges in integrating long-range dependencies to correct locally misaligned features, which results in spatial-spectral distortions. Moreover, these methods tend to be computationally expensive. To address these challenges, we propose a novel detail injection algorithm and develop the invertible attention-guided adaptive convolution and dual-domain Transformer (IACDT) network. In IACDT, we designed an invertible attention mechanism embedded with spectral-spatial attention to efficiently and losslessly extract locally spatial-spectral-aware detail information. In addition, we presented a frequency-spatial dual-domain attention mechanism that combines a frequency-enhanced Transformer and a spatial window Transformer for long-range contextual detail feature correction. This architecture effectively integrates local detail features with long-range dependencies, enabling the model to correct both local misalignments and global inconsistencies. The final HRMS image is obtained through a reconstruction block that consists of residual multireceptive field attention. Extensive experiments demonstrate that IACDT achieves superior fusion performance, computational efficiency, and outstanding results in downstream tasks compared to state-of-the-art methods.https://ieeexplore.ieee.org/document/10845120/Adaptive convolutiondual-domainpansharpeningtransformer
spellingShingle Qun Song
Hangyuan Lu
Chang Xu
Rixian Liu
Weiguo Wan
Wei Tu
Invertible Attention-Guided Adaptive Convolution and Dual-Domain Transformer for Pansharpening
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Adaptive convolution
dual-domain
pansharpening
transformer
title Invertible Attention-Guided Adaptive Convolution and Dual-Domain Transformer for Pansharpening
title_full Invertible Attention-Guided Adaptive Convolution and Dual-Domain Transformer for Pansharpening
title_fullStr Invertible Attention-Guided Adaptive Convolution and Dual-Domain Transformer for Pansharpening
title_full_unstemmed Invertible Attention-Guided Adaptive Convolution and Dual-Domain Transformer for Pansharpening
title_short Invertible Attention-Guided Adaptive Convolution and Dual-Domain Transformer for Pansharpening
title_sort invertible attention guided adaptive convolution and dual domain transformer for pansharpening
topic Adaptive convolution
dual-domain
pansharpening
transformer
url https://ieeexplore.ieee.org/document/10845120/
work_keys_str_mv AT qunsong invertibleattentionguidedadaptiveconvolutionanddualdomaintransformerforpansharpening
AT hangyuanlu invertibleattentionguidedadaptiveconvolutionanddualdomaintransformerforpansharpening
AT changxu invertibleattentionguidedadaptiveconvolutionanddualdomaintransformerforpansharpening
AT rixianliu invertibleattentionguidedadaptiveconvolutionanddualdomaintransformerforpansharpening
AT weiguowan invertibleattentionguidedadaptiveconvolutionanddualdomaintransformerforpansharpening
AT weitu invertibleattentionguidedadaptiveconvolutionanddualdomaintransformerforpansharpening