RemoteCLIP

Links satellite imagery with natural-language descriptions, allowing images to be searched or classified without task-specific training.

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

RemoteCLIP scores how well an image matches a text prompt. It is CLIP fine-tuned on remote sensing imagery, so prompts such as “flooded farmland” or “dense urban” work without task-specific labeling.

In testing across 112 tiles, its scores tracked independently measured cloud cover. It suits narrowing a large archive to the scenes worth examining, and searching imagery collections with poor metadata.

Upstream project →

Catalog entry last checked 2026-08-17. All models →

Run RemoteCLIP 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