ITcon Vol. 30, pg. 852-874, http://www.itcon.org/2025/35

Autonomous Mixed Reality Framework for Real-Time Construction Inspection

DOI:10.36680/j.itcon.2025.035
submitted:November 2024
revised:May 2025
published:May 2025
editor(s):Robert Amor
authors:Tao Boan
Institute for Infrastructure and Environment, School of Engineering, University of Edinburgh, Edinburgh, Scotland, UK
email: boan.tao@ed.ac.uk

Li Jiajun
Institute for Infrastructure and Environment, School of Engineering, University of Edinburgh, Edinburgh, Scotland, UK
email: jiajun.li@ed.ac.uk

Frédéric Bosché
Institute for Infrastructure and Environment, School of Engineering, University of Edinburgh, Edinburgh, Scotland, UK
email: f.bosche@ed.ac.uk
summary:The increasing complexity in construction projects necessitates advancements in the precision and efficiency of inspection processes. In response to this challenge, the present study explores the feasibility of a framework for autonomous inspection using Mixed Reality (MR), Building Information Modelling (BIM) and Artificial Intelligence (AI).The proposed framework encompasses techniques for: object detection in images taken through an MR headset; matching to the object instance in the digital twin; and visualisation of detection results in the MR headset to enable real-time human-in-the-loop decision making, thereby optimising the inspection workflow. The framework's efficacy is evaluated with two datasets representing diverse construction settings, including residential and office environments, focusing on the checking of the presence of ubiquitous elements like electrical sockets and switches. These tests illustrate the practical applicability and limitations of the proposed method.
keywords:Mixed Reality, BIM, Digital Twin, Construction Inspection, Progress Monitoring, Quality Control
full text: (PDF file, 1.478 MB)
citation:Tao B, Li J & Bosché F (2025). Autonomous Mixed Reality Framework for Real-Time Construction Inspection, ITcon Vol. 30, pg. 852-874, https://doi.org/10.36680/j.itcon.2025.035
statistics: