Development of Point Cloud Recognition with Mixed Reality Technology for Automated Quality Control in Light Steel Framing Assembly

This thesis introduces an intelligent tool for construction quality control using 3D point cloud recognition, artificial intelligence, and BIM-based mixed reality. It aims to improve the accuracy and efficiency of quality control of light gauge steel framing (LGSF) system in industrialised construction. By using synthetic BIM data and as-built mixed reality captures, neural networks were trained to identify the types of structural elements accurately. The output is then used to determine the position and dimension of each structural element. The practical value lies in automating construction quality control, which enhances progress monitoring and supports better decision-making, contributing to the digitisation of the construction industry.