CROMA

Creates embeddings from paired radar and optical satellite imagery.

Field
Remote sensing
Method
Foundation model
Data it takes
imagery
License
MIT

CROMA was pretrained on paired Sentinel-1 and Sentinel-2 imagery, learning the relationship between radar and optical views of the same ground. It returns embeddings.

Radar penetrates cloud; optical carries spectral detail. A model trained on both can draw on whichever is available for a given scene.

Runs on CPU.

Upstream project →

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

Run CROMA 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