Faster, Cheaper 3D City Modeling Using Gaussian Splatting

Researchers at the University of Waterloo have developed a faster and cheaper method for creating large-scale 3D models of urban areas. Their system uses Gaussian splatting to automatically generate realistic 3D models from 2D aerial photographs, like those from Google Earth, eliminating the time-consuming manual process. This technology has applications in urban planning, architectural design, and filmmaking, offering high-quality city environments for movies and aiding urban planners in development projects. The team is exploring integrating geospatial AI to expand its capabilities for traffic analysis, solar potential assessment, and more.
Read more