Detecting deviations of building elements with laser scanning

No Thumbnail Available
Journal Title
Journal ISSN
Volume Title
Insinööritieteiden korkeakoulu | Master's thesis
Degree programme
Master's Programme in Geoinformatics (GIS)
The increasing costs and construction delays are common issues during civil engineering projects. These problems are very often related to insufficient geometric quality control of buildings and other civil engineering structures during the construction phase. Traditionally, construction work validation was performed manually, by comparing selective on-site measurements with 2D drawings. In order to identify construction errors more efficiently, researchers proposed to compare as-built state of the construction site with its as-designed model. In construction industry, laser scanning is considered to be a fast and accurate way to continuously gather up-to-date information about situation on the site. The data about the planned state of the project are provided by Building Information Models. In this thesis, Cloud-to-Cloud distance-based method was utilised to compare as-built and as-designed state of the building. From the given Building Information Model and laser scanning point cloud, the proposed approach identified building components deviated from their planned positions. Firstly, the input data were pre-processed and registered in the common coordinate system. Next, Cloud-to-cloud distances were calculated between each planned building element point cloud and laser scanning points. Then, the computed distances were divided into three classes: correct, deviated and missing. The element was assigned a state based on the most frequent class. The implemented method was verified with synthetic as-built point cloud data generated based on Building Information Model. The proposed workflow was tested against data with different deviation values, noise levels and with missing components. The conducted analysis indicated that the proposed method could be a valuable tool for geometric quality control able to identify which components were built incorrectly.
Rönnholm, Petri
Thesis advisor
Kauhanen, Heikki
point cloud, BIM, laser scanning, deviation analysis
Other note