Semi-automated approach to indoor mapping for 3D as-built building information modeling

Abstract BIM (Building Information Modeling), chiefly in the form of a 3D “as-built” model or information-sharing platform, has emerged as a powerful solution to the dynamic requirements of AEC (Architecture, Engineering, and Construction). However, whereas fast and accurate reconstruction of building interiors is essential to any collaborative construction management process, manual creation and utilization of a 3D as-built model typically results in low productivity and erroneous modeling results. This paper proposes a semi-automated method that accounts for and resolves the major problems in the streamlined manual process of 3D as-built model creation in BIM. The method generates a 3D wireframe model combined with clutter data, which is then imported into a BIM tool for as-built modeling. The present study evaluates the proposed method by applying it to two typical rooms in a test building. The 3D as-built model was then subjected to an accuracy assessment using reference points acquired by a total station. The contributions of the proposed method, as compared with fully manual operation, are: (1) reduction of the huge data size of the original point clouds; (2) improvement of the productivity of as-built BIM creation with the aid of 3D wireframes; and (3) accuracy enhancement through a refinement process that entails segmentation and regularization. However, the proposed method is limited to building interiors consisting of planar structures; the modeling of detailed objects, such as windows and doors, unfortunately, still requires manual operation. Thus, further research on detail modeling and refinement is necessary in order to increase the method’s automation and enhancement of its overall suitability for mapping complex indoor environments.