fAIr
Map buildings and roads from satellite imagery, with open models trained close to the ground.
fAIr is an open AI-assisted mapping service built for humanitarian and development work. It detects buildings, roads, waterways and trees in satellite or drone imagery, and suggests them for OpenStreetMap or other open data layers. Anyone can train a model suited to their own region.
And where does AI come in?
Computer vision, on open imagery shared on OpenAerialMap. Base models are retrained on local data, and mappers' corrections go back into training to reduce bias.
The AI label
- Model or technique
- Open computer vision models; outside developers can add their own (HOT, HeiGIT and Mozilla AI so far).
- Model published
- Not filled in
- Where it runs
- Not filled in
- Data
- Satellite and drone imagery shared on OpenAerialMap, and open training data produced by the community.
- Role of people
- Mappers' corrections feed back into the model, and the MapSwipe community validates its predictions (47 projects, more than 850,000 results).
- Frugality
- Not filled in
“Not filled in”: MAUB could not check it with the project. Know more? Write to us.
Run this project? Show the “Listed on MAUB” badge
Paste this code on your site: an image and a link to this page. No script, no tracker: MAUB cannot tell who displays it.
<a href="https://maub.subalpine.it/en/projets/fair-hot"><img src="https://maub.subalpine.it/statique/img/badge-recense-en.svg" alt="Listed on MAUB" width="180" height="48"></a>Something wrong or out of date? Write to contact@subalpine.it.
The monthly newsletter
Once a month, what moved on MAUB, in your inbox.