A new technique allows a single phone camera to build accurate 3D maps in real time, turning any smartphone into a 3D scanning device. Previously, accurate real-time 3D reconstruction required specialized hardware or multiple camera setups.

The approach combines computer vision, neural rendering, and SLAM techniques to extract depth information from monocular video. The result is a dense 3D point cloud that updates continuously as the camera moves through a space.

Applications range from AR experiences and interior design to construction monitoring and autonomous navigation. Making 3D mapping accessible from a single phone camera dramatically lowers the barrier to entry for spatial computing applications.

Source