Machine learning just helped researchers create the biggest 3D map of buildings around the world

Researchers at Germany's Technical University of Munich used machine learning algorithms to analyze 800,000 satellite images and create GlobalBuildingAtlas, mapping 2.75 billion buildings across 97 percent of Earth's surface in unprecedented 3D detail. The database reveals that Asia contains half the world's buildings at 1.22 billion structures with the largest total volume, while Africa has 540 million smaller buildings. This tool enables researchers to track urbanization patterns and compare development across regions in ways impossible with traditional 2D maps.