Microsoft GlobalMLBuildingFootprints

Microsoft's GlobalMLBuildingFootprints project offers a massive open dataset of building outlines covering the entire globe. Bing Maps has utilized machine learning and deep learning techniques to detect approximately 1.4 billion buildings from diverse satellite imagery, including contributions from Maxar, Airbus, and IGN France, captured between 2014 and 2024.

This valuable geospatial data is freely available for download and use under the Open Database License (ODbL), making it an excellent resource for researchers, developers, and GIS professionals. The repository provides regular updates, detailing new additions of building footprints and height estimates, along with information on the imagery sources and geographical regions receiving the largest contributions.

The project emphasizes its commitment to expanding coverage and improving data quality, with recent updates adding millions of new footprints and height estimates for various countries. A central dataset-links.csv file provides access to the continuously growing collection, making it a dynamic and comprehensive data hub for global building information.