Hyperspectral Image Super-Resolution under the Guidance of Deep Gradient Information
Hyperspectral image (HSI) super-resolution has gained great attention in remote sensing, due to its effectiveness in enhancing the spatial information of the HSI while preserving the high spectral discriminative ability, without modifying the imagery hardware. In this paper, we proposed a novel HSI...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-06-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/13/12/2382 |
| _version_ | 1797529630791434240 |
|---|---|
| author | Minghua Zhao Jiawei Ning Jing Hu Tingting Li |
| author_facet | Minghua Zhao Jiawei Ning Jing Hu Tingting Li |
| author_sort | Minghua Zhao |
| collection | DOAJ |
| description | Hyperspectral image (HSI) super-resolution has gained great attention in remote sensing, due to its effectiveness in enhancing the spatial information of the HSI while preserving the high spectral discriminative ability, without modifying the imagery hardware. In this paper, we proposed a novel HSI super-resolution method via a gradient-guided residual dense network (G-RDN), in which the spatial gradient is exploited to guide the super-resolution process. Specifically, there are three modules in the super-resolving process. Firstly, the spatial mapping between the low-resolution HSI and the desired high-resolution HSI is learned via a residual dense network. The residual dense network is used to fully exploit the hierarchical features learned from all the convolutional layers. Meanwhile, the gradient detail is extracted via a residual network (ResNet), which is further utilized to guide the super-resolution process. Finally, an empirical weight is set between the fully obtained global hierarchical features and the gradient details. Experimental results and the data analysis on three benchmark datasets with different scaling factors demonstrated that our proposed G-RDN achieved favorable performance. |
| first_indexed | 2024-03-10T10:17:31Z |
| format | Article |
| id | doaj.art-5a35ba130c9446a19fad3446f59de1aa |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-10T10:17:31Z |
| publishDate | 2021-06-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-5a35ba130c9446a19fad3446f59de1aa2023-11-22T00:43:55ZengMDPI AGRemote Sensing2072-42922021-06-011312238210.3390/rs13122382Hyperspectral Image Super-Resolution under the Guidance of Deep Gradient InformationMinghua Zhao0Jiawei Ning1Jing Hu2Tingting Li3The Shaanxi Key Laboratory for Network Computing and Security Technology, School of Computer Science and Engineering, Xi’an University of Technology, Xi’an 710048, ChinaThe Shaanxi Key Laboratory for Network Computing and Security Technology, School of Computer Science and Engineering, Xi’an University of Technology, Xi’an 710048, ChinaThe Shaanxi Key Laboratory for Network Computing and Security Technology, School of Computer Science and Engineering, Xi’an University of Technology, Xi’an 710048, ChinaThe Shaanxi Key Laboratory for Network Computing and Security Technology, School of Computer Science and Engineering, Xi’an University of Technology, Xi’an 710048, ChinaHyperspectral image (HSI) super-resolution has gained great attention in remote sensing, due to its effectiveness in enhancing the spatial information of the HSI while preserving the high spectral discriminative ability, without modifying the imagery hardware. In this paper, we proposed a novel HSI super-resolution method via a gradient-guided residual dense network (G-RDN), in which the spatial gradient is exploited to guide the super-resolution process. Specifically, there are three modules in the super-resolving process. Firstly, the spatial mapping between the low-resolution HSI and the desired high-resolution HSI is learned via a residual dense network. The residual dense network is used to fully exploit the hierarchical features learned from all the convolutional layers. Meanwhile, the gradient detail is extracted via a residual network (ResNet), which is further utilized to guide the super-resolution process. Finally, an empirical weight is set between the fully obtained global hierarchical features and the gradient details. Experimental results and the data analysis on three benchmark datasets with different scaling factors demonstrated that our proposed G-RDN achieved favorable performance.https://www.mdpi.com/2072-4292/13/12/2382hyperspectral image (HSI)super-resolution (SR)spatial gradientresidual dense network (RDN) |
| spellingShingle | Minghua Zhao Jiawei Ning Jing Hu Tingting Li Hyperspectral Image Super-Resolution under the Guidance of Deep Gradient Information Remote Sensing hyperspectral image (HSI) super-resolution (SR) spatial gradient residual dense network (RDN) |
| title | Hyperspectral Image Super-Resolution under the Guidance of Deep Gradient Information |
| title_full | Hyperspectral Image Super-Resolution under the Guidance of Deep Gradient Information |
| title_fullStr | Hyperspectral Image Super-Resolution under the Guidance of Deep Gradient Information |
| title_full_unstemmed | Hyperspectral Image Super-Resolution under the Guidance of Deep Gradient Information |
| title_short | Hyperspectral Image Super-Resolution under the Guidance of Deep Gradient Information |
| title_sort | hyperspectral image super resolution under the guidance of deep gradient information |
| topic | hyperspectral image (HSI) super-resolution (SR) spatial gradient residual dense network (RDN) |
| url | https://www.mdpi.com/2072-4292/13/12/2382 |
| work_keys_str_mv | AT minghuazhao hyperspectralimagesuperresolutionundertheguidanceofdeepgradientinformation AT jiaweining hyperspectralimagesuperresolutionundertheguidanceofdeepgradientinformation AT jinghu hyperspectralimagesuperresolutionundertheguidanceofdeepgradientinformation AT tingtingli hyperspectralimagesuperresolutionundertheguidanceofdeepgradientinformation |