The six-month project has received £35,000 from the Construction Scotland Innovation Centre’s i-Con Challenge. It is aimed at the management and monitoring of the construction phase of buildings to improve productivity and quality.
The project seeks to carry out defect detection through a combination of ‘deep learning’ algorithms, traditional image processing techniques and VR interfaces.
The team will trial the VR platform on the University of Strathclyde’s new Learning & Teaching building, which is under construction on the campus.
The project partners are the Scottish government's Building Standards Division, Sublime, Robert Gordon University, Balfour Beatty Construction, BDP, Anomalous Technologies, Scottish Futures Trust Hub-Cos Partnership, Highlands Council and the Construction Scotland Innovation Centre (CSIC).
A large, publicly-available dataset of building defect images will be created, which can be used as a standard by the community for algorithm development, testing and comparison.
Dr Andrew Agapiou, senior lecturer in the Department of Architecture and the lead researcher on the project said: “The ability to accurately detect errors and defects in remote and hard-to-access buildings has significant value for the organisations.
“In many situations, these structures are inspected extensively using digital cameras to capture images which can be analysed offline by experts who study the footage in detail.
“The use of advanced photogrammetry and artificial intelligence technologies will help data gathering and analysis with minimum human interaction with buildings.”
The platform will create a 3D immersive environment of a building aimed at assisting safe operation by limiting the need for quantity surveyors and health & safety inspectors to be physically present. It is seen by the researchers as a step towards addressing the limitations and difficulties faced when undertaking inspections on remote, inaccessible sites.
The same digital platform can provide occupants and the wider community with ways to engage with and contribute to a building programme.
The project proposes a framework for integration of methodologies and tools including VR, and digital photogrammetry to collect real-time data to automated decision making.
The CSIC i-Con Challenge is designed to match those faced with current challenges with those in industry and academia who can offer rapid solutions.