#146 · Primary category: Computer Vision

robosat

aerial-imagery machine-learning openstreetmap satellite-imagery segmentation

Semantic segmentation on aerial and satellite imagery. Extracts features such as: buildings, parking lots, roads, water, clouds

Project last updated:06/29/26

GitHub Stars

2.1K

Forks

385

Contributors

17

License

MIT

Why we included this project

RoboSat is a Python 3 pipeline that takes raw aerial and satellite imagery and turns it into map-ready features: buildings, roads, parking lots, even clouds. The whole workflow is covered, from downloading tiles to training a segmentation model to exporting simplified GeoJSON polygons, so the output lands in a shape GIS and mapping tools can use instead of raw masks. The OpenStreetMap diary posts written by the developers show it doing real work, including extracting building footprints from drone imagery in Tanzania, which makes it more instructive as a reference than most archived codebases. One honest caveat: Mapbox no longer maintains it, so treat it as a proven base to study and adapt rather than something to bet an active production service on.

Articles for this project

No articles for this project yet.

To suggest a topic or contribute an article, contact us.

Related projects in this category