A Novel Spectral-Spatial Attention Network for Zero-Shot Pansharpening

Over the past decades, pansharpening technologies have received much attention due to the spatial detail enhancement they introduce into multispectral (MS) images by referencing them to panchromatic images (PAN). Traditional approaches struggle to capture nonlinear relationships between the original...

Full description

Saved in:
Bibliographic Details
Main Authors: Hailiang Lu, Mercedes E. Paoletti, Juan M. Haut, Sergio Moreno-Alvarez, Guangsheng Chen, Weipeng Jing
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/11075521/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849404397004521472
author Hailiang Lu
Mercedes E. Paoletti
Juan M. Haut
Sergio Moreno-Alvarez
Guangsheng Chen
Weipeng Jing
author_facet Hailiang Lu
Mercedes E. Paoletti
Juan M. Haut
Sergio Moreno-Alvarez
Guangsheng Chen
Weipeng Jing
author_sort Hailiang Lu
collection DOAJ
description Over the past decades, pansharpening technologies have received much attention due to the spatial detail enhancement they introduce into multispectral (MS) images by referencing them to panchromatic images (PAN). Traditional approaches struggle to capture nonlinear relationships between the original MS-PAN pair and the pansharpened MS, whilst deep learning (DL)-based methods introduce new challenges. Unsupervised pansharpening is prone to distortion due to the unavailability of reference images and errors in modeling the degradation process. Supervised methods are trained at reduced resolutions before migrating to full resolution, which lead to undesirable results due to scale variations. Meanwhile, the constructed training data cannot encompass all possible scenes, hindering the reconstruction capability of unknown scenes. To tackle these challenges, we propose a novel zero-shot Pansharpening Network (<monospace>ZSPNet</monospace>), which exclusively conducts training and testing on the target image pair, ensuring the robust performance in unknown scenes. Furthermore, <monospace>ZSPNet</monospace> method can effectively reduce the scale variance due to its tailored network using an appropriate patch size. As a result, by integrating 3-D convolutional neural networks (3DCNN), spatial attention and channel attention, <monospace>ZSPNet</monospace> is capable of accurately reconstructing MS with enhanced spatial resolution. The experiments were conducted on the public dataset PAirMax, which have challenging scenes captured by different sensors. Compared to some state-of-the-art traditional and DL-based methods, <monospace>ZSPNet</monospace> demonstrates superior performance in both quantitative assessments and visual results.
format Article
id doaj-art-65c86b166c3c46f693f4b8d07a899712
institution Kabale University
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-65c86b166c3c46f693f4b8d07a8997122025-08-20T03:37:01ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352025-01-0118198051982010.1109/JSTARS.2025.358724411075521A Novel Spectral-Spatial Attention Network for Zero-Shot PansharpeningHailiang Lu0https://orcid.org/0000-0003-0094-184XMercedes E. Paoletti1https://orcid.org/0000-0003-1030-3729Juan M. Haut2https://orcid.org/0000-0001-6701-961XSergio Moreno-Alvarez3https://orcid.org/0000-0002-1858-9920Guangsheng Chen4https://orcid.org/0009-0006-8983-9698Weipeng Jing5https://orcid.org/0000-0001-7933-6946College of Computer and Control Engineering, Northeast Forestry University, Harbin, ChinaDepartment of Technology of Computers and Communications, University of Extremadura, C&#x00E1;ceres, SpainDepartment of Technology of Computers and Communications, University of Extremadura, C&#x00E1;ceres, SpainDepartment of Computer Systems and Languages, National University of Distance Education, Madrid, SpainCollege of Computer and Control Engineering, Northeast Forestry University, Harbin, ChinaCollege of Computer and Control Engineering, Northeast Forestry University, Harbin, ChinaOver the past decades, pansharpening technologies have received much attention due to the spatial detail enhancement they introduce into multispectral (MS) images by referencing them to panchromatic images (PAN). Traditional approaches struggle to capture nonlinear relationships between the original MS-PAN pair and the pansharpened MS, whilst deep learning (DL)-based methods introduce new challenges. Unsupervised pansharpening is prone to distortion due to the unavailability of reference images and errors in modeling the degradation process. Supervised methods are trained at reduced resolutions before migrating to full resolution, which lead to undesirable results due to scale variations. Meanwhile, the constructed training data cannot encompass all possible scenes, hindering the reconstruction capability of unknown scenes. To tackle these challenges, we propose a novel zero-shot Pansharpening Network (<monospace>ZSPNet</monospace>), which exclusively conducts training and testing on the target image pair, ensuring the robust performance in unknown scenes. Furthermore, <monospace>ZSPNet</monospace> method can effectively reduce the scale variance due to its tailored network using an appropriate patch size. As a result, by integrating 3-D convolutional neural networks (3DCNN), spatial attention and channel attention, <monospace>ZSPNet</monospace> is capable of accurately reconstructing MS with enhanced spatial resolution. The experiments were conducted on the public dataset PAirMax, which have challenging scenes captured by different sensors. Compared to some state-of-the-art traditional and DL-based methods, <monospace>ZSPNet</monospace> demonstrates superior performance in both quantitative assessments and visual results.https://ieeexplore.ieee.org/document/11075521/3DCNNchannel attentionpansharpeningremote sensingspatial attentionzero-shot learning
spellingShingle Hailiang Lu
Mercedes E. Paoletti
Juan M. Haut
Sergio Moreno-Alvarez
Guangsheng Chen
Weipeng Jing
A Novel Spectral-Spatial Attention Network for Zero-Shot Pansharpening
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
3DCNN
channel attention
pansharpening
remote sensing
spatial attention
zero-shot learning
title A Novel Spectral-Spatial Attention Network for Zero-Shot Pansharpening
title_full A Novel Spectral-Spatial Attention Network for Zero-Shot Pansharpening
title_fullStr A Novel Spectral-Spatial Attention Network for Zero-Shot Pansharpening
title_full_unstemmed A Novel Spectral-Spatial Attention Network for Zero-Shot Pansharpening
title_short A Novel Spectral-Spatial Attention Network for Zero-Shot Pansharpening
title_sort novel spectral spatial attention network for zero shot pansharpening
topic 3DCNN
channel attention
pansharpening
remote sensing
spatial attention
zero-shot learning
url https://ieeexplore.ieee.org/document/11075521/
work_keys_str_mv AT hailianglu anovelspectralspatialattentionnetworkforzeroshotpansharpening
AT mercedesepaoletti anovelspectralspatialattentionnetworkforzeroshotpansharpening
AT juanmhaut anovelspectralspatialattentionnetworkforzeroshotpansharpening
AT sergiomorenoalvarez anovelspectralspatialattentionnetworkforzeroshotpansharpening
AT guangshengchen anovelspectralspatialattentionnetworkforzeroshotpansharpening
AT weipengjing anovelspectralspatialattentionnetworkforzeroshotpansharpening
AT hailianglu novelspectralspatialattentionnetworkforzeroshotpansharpening
AT mercedesepaoletti novelspectralspatialattentionnetworkforzeroshotpansharpening
AT juanmhaut novelspectralspatialattentionnetworkforzeroshotpansharpening
AT sergiomorenoalvarez novelspectralspatialattentionnetworkforzeroshotpansharpening
AT guangshengchen novelspectralspatialattentionnetworkforzeroshotpansharpening
AT weipengjing novelspectralspatialattentionnetworkforzeroshotpansharpening