Dual Graph U-Nets for Hyperspectral Image Classification
Graph convolutional neural networks (GCNs) have been widely used in hyperspectral images (HSIs) classification for their superiority in processing non-Euclidean structure data. The performance of GCNs relies on the initial graph structure. Most GCN models only utilize spectral information to constru...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2021-01-01
|
| Series: | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/9511023/ |
| _version_ | 1818585563372978176 |
|---|---|
| author | Fangming Guo Zhongwei Li Ziqi Xin Xue Zhu Leiquan Wang Jie Zhang |
| author_facet | Fangming Guo Zhongwei Li Ziqi Xin Xue Zhu Leiquan Wang Jie Zhang |
| author_sort | Fangming Guo |
| collection | DOAJ |
| description | Graph convolutional neural networks (GCNs) have been widely used in hyperspectral images (HSIs) classification for their superiority in processing non-Euclidean structure data. The performance of GCNs relies on the initial graph structure. Most GCN models only utilize spectral information to construct a graph, which is inaccurate because they fail to take the relationship between adjacent nodes into consideration. In addition, due to the over-smooth phenomenon, most GCN models are shallow and unable to extract effective features. To address these issues, a dual graph u-nets is proposed by integrating spatial graph and spectral graph for HSIs classification, denoted by DGU-HSI. To integration the spectral and spatial information, two graphs are constructed for feature extraction. The spectral graph is created by spectral similarity among all pixels where multirange spectral information is retained, and the spatial graph is constructed by exploiting the neighborhood relationship of the center pixel, which describes spatial information. Then, a dual GCN is utilized to extract spatial and spectral graph features simultaneously. To relieve the over-smooth phenomenon, the graph u-nets structure is applied on constructed spectral and spatial graph to extract effective features. Finally, the extracted spectral and spatial features are fused for classification. Experiments conducted on the public datasets demonstrate the effectiveness of the proposed method on HSIs classification. |
| first_indexed | 2024-12-16T08:39:03Z |
| format | Article |
| id | doaj.art-9389d0cb5f8f41feb429b17b0f75a6e9 |
| institution | Directory Open Access Journal |
| issn | 2151-1535 |
| language | English |
| last_indexed | 2024-12-16T08:39:03Z |
| publishDate | 2021-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| spelling | doaj.art-9389d0cb5f8f41feb429b17b0f75a6e92022-12-21T22:37:43ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing2151-15352021-01-01148160817010.1109/JSTARS.2021.31037449511023Dual Graph U-Nets for Hyperspectral Image ClassificationFangming Guo0https://orcid.org/0000-0002-0093-9398Zhongwei Li1https://orcid.org/0000-0002-3934-9053Ziqi Xin2Xue Zhu3Leiquan Wang4https://orcid.org/0000-0003-4314-0030Jie Zhang5https://orcid.org/0000-0002-2635-7783College of Oceanography and Space Informatics, China University of Petroleum (East China), Qingdao, ChinaCollege of Oceanography and Space Informatics, China University of Petroleum (East China), Qingdao, ChinaCollege of Computer Science and Technology, China University of Petroleum (East China), Qingdao, ChinaCollege of Computer Science and Technology, China University of Petroleum (East China), Qingdao, ChinaCollege of Computer Science and Technology, China University of Petroleum (East China), Qingdao, ChinaCollege of Oceanography and Space Informatics, China University of Petroleum (East China), Qingdao, ChinaGraph convolutional neural networks (GCNs) have been widely used in hyperspectral images (HSIs) classification for their superiority in processing non-Euclidean structure data. The performance of GCNs relies on the initial graph structure. Most GCN models only utilize spectral information to construct a graph, which is inaccurate because they fail to take the relationship between adjacent nodes into consideration. In addition, due to the over-smooth phenomenon, most GCN models are shallow and unable to extract effective features. To address these issues, a dual graph u-nets is proposed by integrating spatial graph and spectral graph for HSIs classification, denoted by DGU-HSI. To integration the spectral and spatial information, two graphs are constructed for feature extraction. The spectral graph is created by spectral similarity among all pixels where multirange spectral information is retained, and the spatial graph is constructed by exploiting the neighborhood relationship of the center pixel, which describes spatial information. Then, a dual GCN is utilized to extract spatial and spectral graph features simultaneously. To relieve the over-smooth phenomenon, the graph u-nets structure is applied on constructed spectral and spatial graph to extract effective features. Finally, the extracted spectral and spatial features are fused for classification. Experiments conducted on the public datasets demonstrate the effectiveness of the proposed method on HSIs classification.https://ieeexplore.ieee.org/document/9511023/Graph convolutional networks(GCN)hyperspectral image (HSI) classificationspectral-spatial fusionspectral-spatial graph |
| spellingShingle | Fangming Guo Zhongwei Li Ziqi Xin Xue Zhu Leiquan Wang Jie Zhang Dual Graph U-Nets for Hyperspectral Image Classification IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing Graph convolutional networks(GCN) hyperspectral image (HSI) classification spectral-spatial fusion spectral-spatial graph |
| title | Dual Graph U-Nets for Hyperspectral Image Classification |
| title_full | Dual Graph U-Nets for Hyperspectral Image Classification |
| title_fullStr | Dual Graph U-Nets for Hyperspectral Image Classification |
| title_full_unstemmed | Dual Graph U-Nets for Hyperspectral Image Classification |
| title_short | Dual Graph U-Nets for Hyperspectral Image Classification |
| title_sort | dual graph u nets for hyperspectral image classification |
| topic | Graph convolutional networks(GCN) hyperspectral image (HSI) classification spectral-spatial fusion spectral-spatial graph |
| url | https://ieeexplore.ieee.org/document/9511023/ |
| work_keys_str_mv | AT fangmingguo dualgraphunetsforhyperspectralimageclassification AT zhongweili dualgraphunetsforhyperspectralimageclassification AT ziqixin dualgraphunetsforhyperspectralimageclassification AT xuezhu dualgraphunetsforhyperspectralimageclassification AT leiquanwang dualgraphunetsforhyperspectralimageclassification AT jiezhang dualgraphunetsforhyperspectralimageclassification |