Satlas

Pretrained backbones for satellite and aerial imagery, used as a starting point for task-specific models.

Field
Remote sensing
Method
Foundation model
Data it takes
imagery
License
Apache-2.0

Satlas provides backbones pretrained on large volumes of satellite and aerial imagery, covering Sentinel-2 RGB and multispectral, SAR, and Landsat.

A task head is attached and trained on top, rather than training a full model from scratch.

Clay and Prithvi-EO address the same need. Relative performance depends on sensor and task, so comparing more than one is often worthwhile.

Upstream project →

Catalog entry last checked 2026-08-18. All models →

Run Satlas on your data

The lab runs this on its own compute, in a container, with the inputs and parameters recorded alongside the result. Initial scoping conversations are free.

Contact the Lab