Super-résolution non supervisée d'images hyperspectrales de télédétection utilisant un entraînement entièrement synthétique
Fuente:
arXiv
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908811065819136 |
|---|---|
| author | Xu, Xinxin Gousseau, Yann Kervazo, Christophe Ladjal, Saïd |
| author_facet | Xu, Xinxin Gousseau, Yann Kervazo, Christophe Ladjal, Saïd |
| contents | Hyperspectral single image super-resolution (SISR) aims to enhance spatial resolution while preserving the rich spectral information of hyperspectral images. Most existing methods rely on supervised learning with high-resolution ground truth data, which is often unavailable in practice. To overcome this limitation, we propose an unsupervised learning approach based on synthetic abundance data. The hyperspectral image is first decomposed into endmembers and abundance maps through hyperspectral unmixing. A neural network is then trained to super-resolve these maps using data generated with the dead leaves model, which replicates the statistical properties of real abundances. The final super-resolution hyperspectral image is reconstructed by recombining the super-resolved abundance maps with the endmembers. Experimental results demonstrate the effectiveness of our method and the relevance of synthetic data for training. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_02552 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Super-résolution non supervisée d'images hyperspectrales de télédétection utilisant un entraînement entièrement synthétique Xu, Xinxin Gousseau, Yann Kervazo, Christophe Ladjal, Saïd Image and Video Processing Computer Vision and Pattern Recognition Hyperspectral single image super-resolution (SISR) aims to enhance spatial resolution while preserving the rich spectral information of hyperspectral images. Most existing methods rely on supervised learning with high-resolution ground truth data, which is often unavailable in practice. To overcome this limitation, we propose an unsupervised learning approach based on synthetic abundance data. The hyperspectral image is first decomposed into endmembers and abundance maps through hyperspectral unmixing. A neural network is then trained to super-resolve these maps using data generated with the dead leaves model, which replicates the statistical properties of real abundances. The final super-resolution hyperspectral image is reconstructed by recombining the super-resolved abundance maps with the endmembers. Experimental results demonstrate the effectiveness of our method and the relevance of synthetic data for training. |
| title | Super-résolution non supervisée d'images hyperspectrales de télédétection utilisant un entraînement entièrement synthétique |
| topic | Image and Video Processing Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2602.02552 |