DeepOSM
Train neural networks with OpenStreetMap data and satellite imagery to classify roads and map features.
What is DeepOSM?
DeepOSM is an open-source tool that trains deep learning models to classify roads and other features in satellite imagery, using OpenStreetMap data as ground truth labels. It automates the process of downloading geospatial data, generating training sets, and running neural network training to identify discrepancies or extract features. The project solves the problem of manually verifying and updating map data by providing an automated, scalable computer vision pipeline.
Geospatial developers, data scientists, and OpenStreetMap contributors interested in applying machine learning to map validation, feature extraction, or satellite imagery analysis. It's also relevant for researchers exploring deep learning for remote sensing.
Developers choose DeepOSM because it provides a complete, ready-to-run pipeline that leverages free, open data sources (OSM and NAIP imagery) to train custom models. Its Docker-based setup lowers the barrier to entry, and its focus on practical output—like visualizations of prediction errors—makes it useful for real-world map quality assessment.
Overview
Train a deep learning net with OpenStreetMap features and satellite imagery.
Use Cases
Best For
Related Projects
Found a gem we're missing?
Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.
