Global–Local Multigranularity Transformer for Hyperspectral Image Classification

Hyperspectral image (HSI) classification is a challenging task in remote sensing applications, aiming to determine the category of each pixel by utilizing rich spectral and spatial information in HSI. Convolutional neural networks (CNNs) have been effective in processing HSI data by extracting local...

Full description

Saved in:
Bibliographic Details
Main Authors: Zhe Meng, Qian Yan, Feng Zhao, Gaige Chen, Wenqiang Hua, Miaomiao Liang
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/10746388/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850267897444696064
author Zhe Meng
Qian Yan
Feng Zhao
Gaige Chen
Wenqiang Hua
Miaomiao Liang
author_facet Zhe Meng
Qian Yan
Feng Zhao
Gaige Chen
Wenqiang Hua
Miaomiao Liang
author_sort Zhe Meng
collection DOAJ
description Hyperspectral image (HSI) classification is a challenging task in remote sensing applications, aiming to determine the category of each pixel by utilizing rich spectral and spatial information in HSI. Convolutional neural networks (CNNs) have been effective in processing HSI data by extracting local features, but they are deficient in capturing global contextual information. Recently, transformer has become proficient in attending to global information due to their self-attention mechanisms, yet they may fall short in capturing multiscale features of HSI. To address these limitations, a global–local multigranularity transformer (GLMGT) network is proposed for HSI classification. The GLMGT combines CNN with the transformer to comprehensively capture multigranularity spectral and spatial features across global and local scales. Specifically, we introduce a multigranularity spatial feature extraction block to extensively extract spatial information at different granularities, including multiscale local spatial features and global spatial features. In addition, we introduce a multigranularity spectral feature extraction block to fully leverage spectral information across different granularities. The validity of the proposed method is demonstrated through experimental validation using seven publicly available datasets, which include two Chinese satellite hyperspectral datasets (ZY1-02D Huanghekou and GF-5 Yancheng) and one UAV-based hyperspectral dataset.
format Article
id doaj-art-e87fcdd975e74ecbbcd27b68cfc077c1
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-e87fcdd975e74ecbbcd27b68cfc077c12025-08-20T01:53:36ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352025-01-011811213110.1109/JSTARS.2024.349129410746388Global–Local Multigranularity Transformer for Hyperspectral Image ClassificationZhe Meng0https://orcid.org/0000-0002-2364-2749Qian Yan1https://orcid.org/0009-0006-3425-211XFeng Zhao2https://orcid.org/0000-0002-0323-9573Gaige Chen3https://orcid.org/0000-0002-3045-6514Wenqiang Hua4https://orcid.org/0000-0003-2611-6194Miaomiao Liang5https://orcid.org/0000-0002-4289-7114School of Communications and Information Engineering and School of Artificial Intelligence, Xi'an University of Posts and Telecommunications, Xi'an, ChinaSchool of Communications and Information Engineering and School of Artificial Intelligence, Xi'an University of Posts and Telecommunications, Xi'an, ChinaSchool of Communications and Information Engineering and School of Artificial Intelligence, Xi'an University of Posts and Telecommunications, Xi'an, ChinaSchool of Communications and Information Engineering and School of Artificial Intelligence, Xi'an University of Posts and Telecommunications, Xi'an, ChinaSchool of Computer Science and Technology, Xi'an University of Posts and Telecommunications, Xi'an, ChinaJiangxi Provincial Key Laboratory of Multidimensional Intelligent Perception and Control, School of Information Engineering, Jiangxi University of Science and Technology, Ganzhou, ChinaHyperspectral image (HSI) classification is a challenging task in remote sensing applications, aiming to determine the category of each pixel by utilizing rich spectral and spatial information in HSI. Convolutional neural networks (CNNs) have been effective in processing HSI data by extracting local features, but they are deficient in capturing global contextual information. Recently, transformer has become proficient in attending to global information due to their self-attention mechanisms, yet they may fall short in capturing multiscale features of HSI. To address these limitations, a global–local multigranularity transformer (GLMGT) network is proposed for HSI classification. The GLMGT combines CNN with the transformer to comprehensively capture multigranularity spectral and spatial features across global and local scales. Specifically, we introduce a multigranularity spatial feature extraction block to extensively extract spatial information at different granularities, including multiscale local spatial features and global spatial features. In addition, we introduce a multigranularity spectral feature extraction block to fully leverage spectral information across different granularities. The validity of the proposed method is demonstrated through experimental validation using seven publicly available datasets, which include two Chinese satellite hyperspectral datasets (ZY1-02D Huanghekou and GF-5 Yancheng) and one UAV-based hyperspectral dataset.https://ieeexplore.ieee.org/document/10746388/CNNmultigranularitytransformerhyperspectral image (HSI) classification
spellingShingle Zhe Meng
Qian Yan
Feng Zhao
Gaige Chen
Wenqiang Hua
Miaomiao Liang
Global–Local Multigranularity Transformer for Hyperspectral Image Classification
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
CNN
multigranularity
transformer
hyperspectral image (HSI) classification
title Global–Local Multigranularity Transformer for Hyperspectral Image Classification
title_full Global–Local Multigranularity Transformer for Hyperspectral Image Classification
title_fullStr Global–Local Multigranularity Transformer for Hyperspectral Image Classification
title_full_unstemmed Global–Local Multigranularity Transformer for Hyperspectral Image Classification
title_short Global–Local Multigranularity Transformer for Hyperspectral Image Classification
title_sort global x2013 local multigranularity transformer for hyperspectral image classification
topic CNN
multigranularity
transformer
hyperspectral image (HSI) classification
url https://ieeexplore.ieee.org/document/10746388/
work_keys_str_mv AT zhemeng globalx2013localmultigranularitytransformerforhyperspectralimageclassification
AT qianyan globalx2013localmultigranularitytransformerforhyperspectralimageclassification
AT fengzhao globalx2013localmultigranularitytransformerforhyperspectralimageclassification
AT gaigechen globalx2013localmultigranularitytransformerforhyperspectralimageclassification
AT wenqianghua globalx2013localmultigranularitytransformerforhyperspectralimageclassification
AT miaomiaoliang globalx2013localmultigranularitytransformerforhyperspectralimageclassification