GlobalBuildingAtlas maps 2.75 billion buildings in 3D, enabling new global urban insights

A newly released global 3D map called GlobalBuildingAtlas charts the locations and sizes of about 2.75 billion buildings, a dataset created by researchers at the Technical University of Munich. The atlas blends nearly 800,000 satellite images captured in 2019 with an AI model to estimate building heights where direct measurements are missing. Lead author Professor Xiaoxiang Zhu says the work advances urban analysis by showing not only building footprints but also volumes, which can improve assessments of living conditions and poverty. ”3D building information provides a much more accurate picture of urbanization and poverty than traditional 2D maps,” Zhu said. “With 3D models, we see not only the footprint but also the volume of each building, enabling far more precise insights into living conditions.”

The atlas combines laser-scanning data with artificial intelligence to fill out gaps in the map. The result is a global 3D layer that is exceptionally detailed in urban areas, with a resolution about 30 times finer than existing databases. The publicly accessible map lets users search for an address to view a 3D model of that area, based on data recorded by orbiting satellites. The project launched recently and has drawn more than 280,000 visits, though some visitors report slower loading times due to high demand.

Geographic and demographic breakdowns underscore the global scale. Asia accounts for about 1.22 billion buildings, nearly half of the world’s total, followed by Africa with roughly 540 million and Europe with about 403 million. North America has around 295 million, South America about 264 million, and Oceania roughly 14 million. The researchers also introduce building volume per capita as a proxy for development, noting Finland leads Europe with more than about 3,900 cubic metres of building volume per person, while Greece records markedly lower levels, reflecting uneven development patterns across the region.

The researchers emphasize that the dataset is not only a curiosity. It is designed as a tool for planning, disaster prevention, and urban development. By revealing how much space people have per person and where dense concentrations of structure exist, officials can better target housing construction, public facilities, and emergency infrastructure. The German Aerospace Center is already exploring how to use GlobalBuildingAtlas as part of its disaster support work around the world.

The map does not merely capture modern skylines. In areas where detectable, the 3D data can show features such as the layout of city walls or the shapes of bridges and boats on rivers, illustrating the breadth of information now within reach. In London, for example, the satellite data used to create the map can echo the silhouette of Canary Wharf’s towers and other major landmarks, while in historic sites the model can reveal the outline of old fortifications and walls. The tool’s high resolution means it can even identify details as fine as the contours of animal enclosures in urban zoos in some datasets, though the primary focus remains on occupied buildings and public spaces.

Despite its promise, the project faces practical limitations. The current database relies on imagery from 2019 and is most precise in densely built environments. The sheer volume of data means the public-facing interface can be slow at peak times, and researchers caution that the dataset will require ongoing validation as new imagery becomes available. Still, proponents argue the GlobalBuildingAtlas marks a significant advance in how scientists and policymakers can quantify urban form at a planetary scale, potentially informing everything from climate resilience planning to housing policy and infrastructure investment.