Gaussian Building Mesh (GBM): Extract a building’s 3D mesh with Google Earth and Gaussian Splatting

The rapid convergence of computer vision and digital technologies is redefining how buildings are captured, modeled, and managed. In computer vision, recently released open-source pre-trained foundational image segmentation and object detection models allow for geometrically consistent segmentation of objects of interest in multi-view 2D images. Text-based or click-based prompts can be used to segment objects of interest without requiring labeled training datasets, allowing for both user-prompted and automated segmentation. Simultaneously, Gaussian Splatting allows for learning a 3D representation of a scene’s geometry and radiance based on 2D images. Combining Google Earth Studio, SAM2+GroundingDINO, 2D Gaussian Splatting, and our improvements in mask refinement based on morphological operations and contour simplification, we created a pipeline to extract the 3D mesh of any building based on its name, address, or geographic coordinates. Our pipeline offers a fast and user-accessible framework for rapid 3D modeling of built environments and structures, enabling downstream applications. • Novel 3D mesh extraction pipeline based on text or click-based user input. • Extract a building 3D mesh from its name, address, or geocoding information. • Uses Google Earth data and does not require any on-site data.