Dual-Branch Network With Mutual Guidance for Hyperspectral Image Superresolution

Fusion-based hyperspectral image superresolution has recently attracted increasing interest due to its superior reconstruction quality. This approach enhances the spatial resolution of low-resolution hyperspectral images (LR-HSIs) by fusing high-resolution multispectral images (HR-MSIs) of the same...

Full description

Saved in:
Bibliographic Details
Main Authors: Tiejun Zhang, Boqun Zhang, Nan Hu, Xinyu Li, Peng Li
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/11002502/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850219313363943424
author Tiejun Zhang
Boqun Zhang
Nan Hu
Xinyu Li
Peng Li
author_facet Tiejun Zhang
Boqun Zhang
Nan Hu
Xinyu Li
Peng Li
author_sort Tiejun Zhang
collection DOAJ
description Fusion-based hyperspectral image superresolution has recently attracted increasing interest due to its superior reconstruction quality. This approach enhances the spatial resolution of low-resolution hyperspectral images (LR-HSIs) by fusing high-resolution multispectral images (HR-MSIs) of the same scene. However, most existing deep learning-based methods have not sufficiently considered the huge modality differences between these two types of images before feature fusion, potentially resulting in the loss of valuable information. To tackle this challenge, we introduce a novel dual-branch network with mutual guidance (DBMGNet). Specifically, we employ a dual-branch architecture to process the input LR-HSI and HR-MSI in parallel, with the aim of reconciling their modality differences. For this purpose, we designed a mutually guided dual-stream transformer block that performs bidirectional calibration between the branches, enhancing their spatial and spectral consistency. To prevent excessive coupling of information, we proposed a multiscale feature enhancement block, which independently refines fine-grained details within each branch. Finally, a weighted feature fusion block is developed to effectively integrate the features from both branches. Experiments on three widely used datasets indicate that the proposed DBMGNet achieves stable and superior performance with lower computational cost in comparison with several state-of-the-art approaches.
format Article
id doaj-art-b9f736d4f7d54066b26a518aacd7caac
institution OA Journals
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-b9f736d4f7d54066b26a518aacd7caac2025-08-20T02:07:24ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352025-01-0118143681438110.1109/JSTARS.2025.356933111002502Dual-Branch Network With Mutual Guidance for Hyperspectral Image SuperresolutionTiejun Zhang0https://orcid.org/0009-0005-9644-2456Boqun Zhang1https://orcid.org/0009-0000-5348-6882Nan Hu2https://orcid.org/0009-0009-7714-6925Xinyu Li3https://orcid.org/0009-0006-7887-2767Peng Li4https://orcid.org/0009-0002-7688-0473College of Computer Science and Software, Harbin University of Science and Technology, Harbin, ChinaCollege of Computer Science and Software, Harbin University of Science and Technology, Harbin, ChinaCollege of Computer Science and Technology, Jilin University, Changchun, ChinaCollege of Computer Science and Software, Harbin University of Science and Technology, Harbin, ChinaCollege of Computer Science and Software, Harbin University of Science and Technology, Harbin, ChinaFusion-based hyperspectral image superresolution has recently attracted increasing interest due to its superior reconstruction quality. This approach enhances the spatial resolution of low-resolution hyperspectral images (LR-HSIs) by fusing high-resolution multispectral images (HR-MSIs) of the same scene. However, most existing deep learning-based methods have not sufficiently considered the huge modality differences between these two types of images before feature fusion, potentially resulting in the loss of valuable information. To tackle this challenge, we introduce a novel dual-branch network with mutual guidance (DBMGNet). Specifically, we employ a dual-branch architecture to process the input LR-HSI and HR-MSI in parallel, with the aim of reconciling their modality differences. For this purpose, we designed a mutually guided dual-stream transformer block that performs bidirectional calibration between the branches, enhancing their spatial and spectral consistency. To prevent excessive coupling of information, we proposed a multiscale feature enhancement block, which independently refines fine-grained details within each branch. Finally, a weighted feature fusion block is developed to effectively integrate the features from both branches. Experiments on three widely used datasets indicate that the proposed DBMGNet achieves stable and superior performance with lower computational cost in comparison with several state-of-the-art approaches.https://ieeexplore.ieee.org/document/11002502/Dual-branch networkhyperspectral image superresolutionimage fusiontransformer
spellingShingle Tiejun Zhang
Boqun Zhang
Nan Hu
Xinyu Li
Peng Li
Dual-Branch Network With Mutual Guidance for Hyperspectral Image Superresolution
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Dual-branch network
hyperspectral image superresolution
image fusion
transformer
title Dual-Branch Network With Mutual Guidance for Hyperspectral Image Superresolution
title_full Dual-Branch Network With Mutual Guidance for Hyperspectral Image Superresolution
title_fullStr Dual-Branch Network With Mutual Guidance for Hyperspectral Image Superresolution
title_full_unstemmed Dual-Branch Network With Mutual Guidance for Hyperspectral Image Superresolution
title_short Dual-Branch Network With Mutual Guidance for Hyperspectral Image Superresolution
title_sort dual branch network with mutual guidance for hyperspectral image superresolution
topic Dual-branch network
hyperspectral image superresolution
image fusion
transformer
url https://ieeexplore.ieee.org/document/11002502/
work_keys_str_mv AT tiejunzhang dualbranchnetworkwithmutualguidanceforhyperspectralimagesuperresolution
AT boqunzhang dualbranchnetworkwithmutualguidanceforhyperspectralimagesuperresolution
AT nanhu dualbranchnetworkwithmutualguidanceforhyperspectralimagesuperresolution
AT xinyuli dualbranchnetworkwithmutualguidanceforhyperspectralimagesuperresolution
AT pengli dualbranchnetworkwithmutualguidanceforhyperspectralimagesuperresolution