Planning for scanning in construction : optimizing 3D laser scanning operations using building information modelling and a novel specification on surface scanning completeness.

Abstract

Application of Terrestrial Laser Scanning (TLS) technology in the Architectural Engineering and Construction (AEC) industry is gaining popularity because the technology uniquely offers the means to create as-built three-dimensional (3D) models of existing facilities, and conduct construction project progress and dimensional quality measurements. An open challenge with regard to the use TLS for such applications is to efficiently generate effective scanning plans that satisfy pre-defined point cloud quality specifications. Two such specifications are currently commonly used: Level of Accuracy (LOA) that focuses on individual point precision, and Level of Detail (LOD) that focuses on point density. Given such specifications, current practice sees professionals manually prepare scanning plans using existing 2D CAD drawings, some ad-hoc rules (of thumb), and their experience. Yet, it is difficult to manually generate and analyse laser scanning plans to ensure they satisfy scanning quality specifications such as those above. Manually-defined plans may easily lead to over-scanning, or worse under-scanning with incomplete data (which may require the team to go back on site to acquire complementary data). To minimize the risk of producing inadequate scanning plans, some semi-automated and automated methods have been proposed by researchers that use the 3D (BIM) model generated during the design stage. These methods take consideration for LOA and LOD. However, these are point-based specifications that do not guarantee that a sufficient amount of the surface of each object is covered by the acquired data, despite this aspect being important to many of the applications for which TLS is employed (e.g. modelling existing facilities). Therefore, this research uniquely proposes a novel planning for scanning quality specification, called Level of Surface Completeness (LOC) that assesses point cloud quality in terms of surface completeness. In addition, an approach is proposed for automatic planning for scanning in the AEC industry that takes both LOA and LOC specifications into account. The approach is ‘generic’ in the sense that it can be employed for any type of project. It is designed to generate automatic laser scanning plans using as input: (1) the facility’s 3D BIM model; (2) the scanner’s characteristics; and (3) the LOA and LOC specifications. The output is the smallest set of scanning locations necessary to achieve those requirements. The optimal solution is found by formulating the problem as a binary integer programming optimization problem, which is easily solved using a branch-and-cut algorithm. To assess the performance of the approach, experiments are conducted using a simple concrete structural model, a more complex structural model, and a section of the latter extended with Mechanical Electrical and Plumbing (MEP) components. The overall performance of the proposed approach for automatic planning for scanning is encouraging, showing that it is possible to take surface-based specifications into account in automated planning-for-scanning algorithms. However, the experimental results also highlight a significant weakness of the approach presented here which is that it does not take into account the overlapping of surfaces covered from different scanning locations and thus may inaccurately assess covered surfaces.