Saved in:
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2512.01181 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866912740187045888 |
|---|---|
| author | Du, Andrew Del Prete, Roberto Mousist, Alejandro Manser, Nick Marre, Fabrice Barton, Andrew Seubert, Carl Meoni, Gabriele Chin, Tat-Jun |
| author_facet | Du, Andrew Del Prete, Roberto Mousist, Alejandro Manser, Nick Marre, Fabrice Barton, Andrew Seubert, Carl Meoni, Gabriele Chin, Tat-Jun |
| contents | Geospatial foundation models (GeoFMs) promise broad generalisation capacity for Earth observation (EO) tasks, particularly under data-limited conditions. However, their large size poses a barrier to deployment on resource-constrained space hardware. To address this, we present compact variants of a Vision Transformer (ViT)-based GeoFM that preserve downstream task performance while enabling onboard execution. Evaluation across five downstream tasks and validation in two representative flight environments show that model compression and domain adaptation are critical to reducing size and resource demands while maintaining high performance under operational conditions. We further demonstrate reliable on-orbit inference with the IMAGIN-e payload aboard the International Space Station. These results establish a pathway from large GeoFMs to flight-ready, resource-efficient deployments, expanding the feasibility of onboard AI for EO missions. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_01181 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | First On-Orbit Demonstration of a Geospatial Foundation Model Du, Andrew Del Prete, Roberto Mousist, Alejandro Manser, Nick Marre, Fabrice Barton, Andrew Seubert, Carl Meoni, Gabriele Chin, Tat-Jun Machine Learning Artificial Intelligence Computer Vision and Pattern Recognition Geospatial foundation models (GeoFMs) promise broad generalisation capacity for Earth observation (EO) tasks, particularly under data-limited conditions. However, their large size poses a barrier to deployment on resource-constrained space hardware. To address this, we present compact variants of a Vision Transformer (ViT)-based GeoFM that preserve downstream task performance while enabling onboard execution. Evaluation across five downstream tasks and validation in two representative flight environments show that model compression and domain adaptation are critical to reducing size and resource demands while maintaining high performance under operational conditions. We further demonstrate reliable on-orbit inference with the IMAGIN-e payload aboard the International Space Station. These results establish a pathway from large GeoFMs to flight-ready, resource-efficient deployments, expanding the feasibility of onboard AI for EO missions. |
| title | First On-Orbit Demonstration of a Geospatial Foundation Model |
| topic | Machine Learning Artificial Intelligence Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2512.01181 |