3D Reconstruction of Interior Wall Surfaces under Occlusion and Clutter

Abstract

Laser scanners are often used to create 3D models of buildings for civil engineering applications. The current manual process is time-consuming and error-prone. This paper presents a method for using laser scanner data to model predominantly planar surfaces, such as walls, floors, and ceilings, despite the presence of significant amounts of clutter and occlusion, which occur frequently in natural indoor environments. Our goal is to recover the surface shape, detect and model any openings, and fill in the occluded regions. Our method identifies candidate surfaces for modeling, labels occluded surface regions, detects openings in each surface using supervised learning, and reconstructs the surface in the occluded regions. We evaluate the method on a large, highly cluttered data set of a building consisting of forty separate rooms.