Abstract
Building Information Modelling (BIM) has recently been widely applied in the Architecture, Engineering, and Construction Industry (AEC). BIM graphical information can provide a more intuitive display of the building and its content. Among all types of BIM, Historical Building Information Modelling (HBIM) focuses on creating BIM for historical buildings, which can be used for the documentation and preservation of these buildings. However, because historical buildings were built before the BIM era, many of these buildings lack an HBIM. To create an HBIM model for these buildings, Terrestrial Laser Scanning (TLS) and photogrammetry are commonly used to collect point cloud data and provide geometric and semantic information. HBIM has several applications, such as digital documentation, conservation, restoration, maintenance, etc, of historical buildings. Each of these applications requires accurate and complete data, underscoring the importance of accuracy in HBIM models. According to the Beijing Administration for Market Regulation, the digital documentation of historical buildings normally requires an error of less than 20 mm for the main body of the building and within 2 mm for detailed architectural parts, such as carving and reliefs, etc. In addition, the collected data should cover the whole building, including the structural components and detailed elements, representing the need for a complete dataset. However, the data collected by TLS and photogrammetry may be inaccurate or incomplete, which could lead to inaccuracies and data loss in HBIM models. Therefore, the accuracy and completeness of the data collected by these devices needs to be assessed for the reliable application of HBIM in the preservation and management of historical buildings.While previous studies evaluated the accuracy of the point clouds of exterior heritage sites, there remains a gap in evaluating the accuracy and completeness of indoor heritage clouds. Unlike assessing the point clouds in an outdoor environment, indoor areas may have many obstructions that could block the GNSS signals and lights, therefore making those clouds created in indoor areas different from those outdoors. Consequently, to evaluate the accuracy of the point clouds in indoor areas, this thesis proposes a three-step approach that uses the measurement from a Total Station (TS) as a reference to evaluate the accuracy and completeness of the indoor 3D point clouds of historical buildings. The three steps include: (i) sectional-based visual inspection and cloud density analysis, (ii) cloud to cloud (C2C) comparison, and (iii) point to point (P2P) comparison. The approach can evaluate the completeness and geometric accuracy of the created point clouds, thereby aiding in the creation of more accurate HBIM models. To demonstrate the effectiveness of the proposed approach, a 3D point cloud and a photogrammetry cloud of the interior of an ancient cathedral were used as an exmaple to demonstrate the proposed approach. Results indicate that the TLS cloud has a higher cloud density and completeness, but showed a higher mean P2P value as compared to the photogrammetry cloud. These findings demonstrated the effectiveness of using the proposed approach in evaluating both the completeness and accuracy of point clouds, thus supporting the development of more precise HBIM models for historical buildings. Additionally, since the HBIM model creation process lacks a benchmark to evaluate the point cloud data sources, the proposed workflow would be beneficial for the best practice guidance for evaluating point clouds.
For functional buildings that have BIM models, changes may occur in the building’s content during the Operation and Maintenance (O&M) stage of the lifecycle of a building and cause inaccuracies in the BIM model, which could lead to inappropriate decisions. This thesis proposes an approach based on Pedestrian Dead Reckoning (PDR) for an Inertial Measurement Unit (IMU) integrated with a Mobile Laser Scanner (MLS) to create room-based 3D point clouds. Unlike the conventional use of Terrestrial Laser Scanner (TLS), the proposed approach employs a low-cost Mobile Laser Scanner (MLS) combined with an Inertial Measurement Unit (IMU) for indoor scanning. This method bypasses the need for selecting scanning points and leveling the TLS, leading to a more efficient and cost-effective generation of point clouds. By moving the system in nearly straight lines, multiple indoor 3D point clouds will be created. However, each individual scan is captured within its local coordinate system, necessitating the registration of these clouds to form a unified representation of the indoor environment. To address this, the thesis introduces a novel registration method, which successfully determines the correct spatial alignment and rotational transformation for each scan. Three buildings of varying sizes and shapes were scanned. The results demonstrate that the proposed approach can produce room-based 3D point clouds with centimetre-level accuracy, offering more efficient graphical information for updating BIM models compared to TLS.
| Date of Award | 15 Mar 2025 |
|---|---|
| Original language | English |
| Awarding Institution |
|
| Supervisor | Yung-Tsang Chen (Supervisor), Nicholas Hamm (Supervisor), Zhiang Zhang (Supervisor) & Craig Hancock (Supervisor) |