Simplification of geometric objects in an indoor space

• Published in: ISPRS journal of photogrammetry and remote sensing Vol. 147; pp. 146 - 162
• Main Authors: Kim, Joon-Seok, Li, Ki-Joune
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.01.2019
• Subjects: 3D model; Indoor space; Level of detail (LoD); Prism model; Simplification; 3D model; Prism model; Simplification; Level of detail (LoD); Indoor space
• ISSN: 0924-2716; 1872-8235
• DOI: 10.1016/j.isprsjprs.2018.11.017

The interior of a building may be more complicated than its exterior because such an indoor space comprises a number of three-dimensional (3D) non-overlapping regions called cells (e.g. rooms). Owing to the complexity of 3D geometry, applications with 3D data (e.g. indoor navigation) require an adequate level of detail (LoD) to achieve their purpose. To supply the 3D data demanded by clients, a customised simplification that considers LoDs in an indoor space is required. Most research studies on the simplification of 3D objects, however, have focused on general 3D objects or the exterior of buildings. Applying such approaches to indoor space objects is inefficient and may cause loss of important information because cells in an indoor space have distinctive characteristics compared to general 3D objects. For instance, a conventional room is surrounded by vertically aligned walls and a horizontally aligned ceiling and floor. For this reason, we propose a dedicated simplification method of 3D geometric objects in an indoor space. Our method takes full advantage of the prism model, which is an alternative 3D geometric model that supports prismatic shapes motivated by traits of indoor spaces. Additionally, an approach for dealing with potential topological inconsistencies during simplification is presented in this paper. An empirical analysis of the efficiency of the proposed simplification is conducted to validate our work.