An Improved Hyperspectral Unmixing Approach Based on a Spatial–Spectral Adaptive Nonlinear Unmixing Network
The autoencoder (AE) framework is usually adopted as a baseline network for hyperspectral unmixing. Totally an AE performs well in hyperspectral unmixing through automatically learning low-dimensional embedding and reconstructing data. However, most available AE-based hyperspectral unmixing networks...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2023-01-01
|
| Series: | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10278425/ |
| _version_ | 1797635433341911040 |
|---|---|
| author | Xiao Chen Xianfeng Zhang Miao Ren Bo Zhou Ziyuan Feng Junyi Cheng |
| author_facet | Xiao Chen Xianfeng Zhang Miao Ren Bo Zhou Ziyuan Feng Junyi Cheng |
| author_sort | Xiao Chen |
| collection | DOAJ |
| description | The autoencoder (AE) framework is usually adopted as a baseline network for hyperspectral unmixing. Totally an AE performs well in hyperspectral unmixing through automatically learning low-dimensional embedding and reconstructing data. However, most available AE-based hyperspectral unmixing networks do not fully consider the spatial and spectral information of different ground features in hyperspectral images and output relatively fixed ratios of linear and nonlinear photon scattering effects under different scenarios. Therefore, these methods have poor generalization abilities across different ground features and scenarios. Here, inspired by the two-stream network structure, we propose a spatial–spectral adaptive nonlinear unmixing network (SSANU-Net) in which the spatial–spectral information of hyperspectral imagery is effectively learned using the two-stream encoder, followed by the simulation of the linear–nonlinear scattering component of photons using a two-stream decoder. Additionally, we adopt a combination of spatial–spectral and linear–nonlinear components using the optimized adaptive weighting strategy of learnable parameters. Experiments with several hyperspectral image datasets (i.e., Samson, Jasper Ridge, and Urban) showed that the proposed SSANU-Net network had higher unmixing accuracy and generalization performance compared with several conventional methods. This demonstrates that SSANU-Net represents a novel method for hyperspectral unmixing analysis. |
| first_indexed | 2024-03-11T12:22:04Z |
| format | Article |
| id | doaj.art-badcd2458b5348369e4d093fa629241b |
| institution | Directory Open Access Journal |
| issn | 2151-1535 |
| language | English |
| last_indexed | 2024-03-11T12:22:04Z |
| publishDate | 2023-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| spelling | doaj.art-badcd2458b5348369e4d093fa629241b2023-11-07T00:00:29ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing2151-15352023-01-01169680969610.1109/JSTARS.2023.332374810278425An Improved Hyperspectral Unmixing Approach Based on a Spatial–Spectral Adaptive Nonlinear Unmixing NetworkXiao Chen0https://orcid.org/0000-0001-9221-1603Xianfeng Zhang1https://orcid.org/0000-0002-2475-4558Miao Ren2Bo Zhou3https://orcid.org/0009-0009-6123-1332Ziyuan Feng4Junyi Cheng5https://orcid.org/0000-0002-9225-2824Institute of Remote Sensing and Geographic Information, Peking University, Beijing, ChinaInstitute of Remote Sensing and Geographic Information, Peking University, Beijing, ChinaInstitute of Remote Sensing and Geographic Information, Peking University, Beijing, ChinaInstitute of Remote Sensing and Geographic Information, Peking University, Beijing, ChinaInstitute of Remote Sensing and Geographic Information, Peking University, Beijing, ChinaInstitute of Remote Sensing and Geographic Information, Peking University, Beijing, ChinaThe autoencoder (AE) framework is usually adopted as a baseline network for hyperspectral unmixing. Totally an AE performs well in hyperspectral unmixing through automatically learning low-dimensional embedding and reconstructing data. However, most available AE-based hyperspectral unmixing networks do not fully consider the spatial and spectral information of different ground features in hyperspectral images and output relatively fixed ratios of linear and nonlinear photon scattering effects under different scenarios. Therefore, these methods have poor generalization abilities across different ground features and scenarios. Here, inspired by the two-stream network structure, we propose a spatial–spectral adaptive nonlinear unmixing network (SSANU-Net) in which the spatial–spectral information of hyperspectral imagery is effectively learned using the two-stream encoder, followed by the simulation of the linear–nonlinear scattering component of photons using a two-stream decoder. Additionally, we adopt a combination of spatial–spectral and linear–nonlinear components using the optimized adaptive weighting strategy of learnable parameters. Experiments with several hyperspectral image datasets (i.e., Samson, Jasper Ridge, and Urban) showed that the proposed SSANU-Net network had higher unmixing accuracy and generalization performance compared with several conventional methods. This demonstrates that SSANU-Net represents a novel method for hyperspectral unmixing analysis.https://ieeexplore.ieee.org/document/10278425/Adaptive weightingautoencoder (AE)hyperspectral imagerynonlinear mixingspatial–spectral adaptive nonlinear unmixing network (SSANU-Net) |
| spellingShingle | Xiao Chen Xianfeng Zhang Miao Ren Bo Zhou Ziyuan Feng Junyi Cheng An Improved Hyperspectral Unmixing Approach Based on a Spatial–Spectral Adaptive Nonlinear Unmixing Network IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing Adaptive weighting autoencoder (AE) hyperspectral imagery nonlinear mixing spatial–spectral adaptive nonlinear unmixing network (SSANU-Net) |
| title | An Improved Hyperspectral Unmixing Approach Based on a Spatial–Spectral Adaptive Nonlinear Unmixing Network |
| title_full | An Improved Hyperspectral Unmixing Approach Based on a Spatial–Spectral Adaptive Nonlinear Unmixing Network |
| title_fullStr | An Improved Hyperspectral Unmixing Approach Based on a Spatial–Spectral Adaptive Nonlinear Unmixing Network |
| title_full_unstemmed | An Improved Hyperspectral Unmixing Approach Based on a Spatial–Spectral Adaptive Nonlinear Unmixing Network |
| title_short | An Improved Hyperspectral Unmixing Approach Based on a Spatial–Spectral Adaptive Nonlinear Unmixing Network |
| title_sort | improved hyperspectral unmixing approach based on a spatial x2013 spectral adaptive nonlinear unmixing network |
| topic | Adaptive weighting autoencoder (AE) hyperspectral imagery nonlinear mixing spatial–spectral adaptive nonlinear unmixing network (SSANU-Net) |
| url | https://ieeexplore.ieee.org/document/10278425/ |
| work_keys_str_mv | AT xiaochen animprovedhyperspectralunmixingapproachbasedonaspatialx2013spectraladaptivenonlinearunmixingnetwork AT xianfengzhang animprovedhyperspectralunmixingapproachbasedonaspatialx2013spectraladaptivenonlinearunmixingnetwork AT miaoren animprovedhyperspectralunmixingapproachbasedonaspatialx2013spectraladaptivenonlinearunmixingnetwork AT bozhou animprovedhyperspectralunmixingapproachbasedonaspatialx2013spectraladaptivenonlinearunmixingnetwork AT ziyuanfeng animprovedhyperspectralunmixingapproachbasedonaspatialx2013spectraladaptivenonlinearunmixingnetwork AT junyicheng animprovedhyperspectralunmixingapproachbasedonaspatialx2013spectraladaptivenonlinearunmixingnetwork AT xiaochen improvedhyperspectralunmixingapproachbasedonaspatialx2013spectraladaptivenonlinearunmixingnetwork AT xianfengzhang improvedhyperspectralunmixingapproachbasedonaspatialx2013spectraladaptivenonlinearunmixingnetwork AT miaoren improvedhyperspectralunmixingapproachbasedonaspatialx2013spectraladaptivenonlinearunmixingnetwork AT bozhou improvedhyperspectralunmixingapproachbasedonaspatialx2013spectraladaptivenonlinearunmixingnetwork AT ziyuanfeng improvedhyperspectralunmixingapproachbasedonaspatialx2013spectraladaptivenonlinearunmixingnetwork AT junyicheng improvedhyperspectralunmixingapproachbasedonaspatialx2013spectraladaptivenonlinearunmixingnetwork |