This Blueprint helps you use object detection and image segmentation models to identify and map features in OpenStreetMap.
Get started right away finding swimming pools and contributing them to OpenStreetMap:
| Google Colab | HuggingFace Spaces | GitHub Codespaces |
|---|---|---|
| Process full Area Around a Point |
You can also create your own dataset and finetune a new model for a different use case:
| Create Dataset | Finetune Model |
|---|---|
This project is licensed under the AGPL-3.0 License. See the LICENSE file for details.
Contributions are welcome! To get started, you can check out the CONTRIBUTING.md file.
This blueprint is a high level guide to feature extraction and preparation of data in OpenStreetMap compatible formats; designed to improve the accessibility of computer vision tooling to mapping users and document potential processes.
It is your responsibility to ensure the following prior to any data additions, as a minimum:
- Adequately document the limitations of your model(s)
- Have performed significant QA and validated the changes are at or above what a typical human mapper would generate
- Are fully aware of Import Guidelines
- Have a plan for granular human-in-the-loop verification, for example use of the Rapid editor or a MapRoulette challenge.
- Have community buy in