Scientists Unveil Unprecedented 3D Map of Every Building on Earth

Scientists Unveil Unprecedented 3D Map of Every Building on Earth

Alex Duffy
Alex Duffy
2 Min.
Aerial view of a city with buildings, roads, trees, and grass, featuring a central map of a proposed development site.

Scientists Unveil Unprecedented 3D Map of Every Building on Earth

Scientists Create 3D Map of Nearly Every Building on Earth

There are 2.75 billion of them.

Scientists have produced the most comprehensive three-dimensional map of global urban development to date. Dubbed the Global Building Atlas, the project covers roughly 97% of the world's buildings—some 2.75 billion structures.

The dataset was published in the open-access journal Earth System Science Data. Combining satellite imagery with machine learning, the map displays building footprints and heights with a spatial resolution of 3 by 3 meters.

Xiaoxiang Zhu, a remote sensing expert at the Technical University of Munich and co-author of the study, explained that the new 3D map unlocks vast potential for urban planning analysis, disaster risk assessment, and climate change modeling. It could also aid in tracking progress toward the UN's Sustainable Development Goals related to cities and human settlements.

To build the map, researchers processed around 800,000 satellite images from 2019. Deep learning algorithms were trained using benchmark LiDAR laser-scanning data collected in 168 cities across Europe, North America, and Oceania. This approach enabled the team to predict building height, volume, and area—even in regions previously lacking detailed 3D data.

Experts note that the Global Building Atlas could become a critical tool for evaluating urban vulnerability to floods, earthquakes, and other disasters, as well as for studying urbanization trends and even detecting potential corruption in construction sectors.

Until now, global urban datasets have largely been limited to two-dimensional maps. Creating a worldwide 3D model had long been considered a daunting challenge due to the high cost of LiDAR scanning and the limited availability of high-resolution imagery.