Hash-Based Method for Generating Building Information Models From 2D CAD Drawings

Abstract

A building information model (BIM) encompasses the essential information required for a building’s entire lifecycle, based on a 3D model. However, because building design begins with spatial planning, the initial design is based on a 2D plane. Furthermore, introducing BIM to general buildings is not effective in terms of modeling because of the design cost and the inherent 2D structure of older buildings. Consequently, research conducted over the past few years aimed to automatically generate 3D models from 2D drawings; however, the performance is limited in terms of data structure and methodology. Therefore, this study proposes a method that generates 3D information necessary for generating BIM models from 2D drawings using 2D hash-based data structures and planar graph analysis to resolve the above problem. Time and spatial complexities for planar graph generation from drawings consisting of n line segments among the proposed methods are $O(kn)$ and $O(n)$ , respectively, where n is the number of the lines in drawing, k satisfies $E(p)\leq k \leq n$ , and $E(p)$ denotes the average number of intersections in each hash cell. By applying it to 2D drawing data used in actual architectures, a 3D model could be developed with approximately 2.9% effort compared to the entire 2D drawing being modeled by a worker. This enables designers unfamiliar with BIM environments to conveniently obtain IFC files containing BIM models and then utilize BIM software to harness 3D information.