Transformation of Point Clouds to Images for Safety Analysis of Scaffold Joints

Structural flaws in scaffolds can lead to fatal accidents on construction sites, and the installation status of scaffold joints is crucial for the safety of scaffolds. However, manual inspection of scaffold joints can be challenging due to their small size and large quantity. This paper presents a scaffold joint inspection mechanism using a Terrestrial Laser Scanner (TLS) to address the challenge of inspecting large numbers of scaffold joints for structural flaws. Scaffold members are first extracted from the TLS-acquired point clouds using a 3D segmentation model to generate scaffold joint image sets. Then, a rule-based classifier is used on the image sets to identify the installation status of each joint. Our experiment showed 82.1% accuracy in scaffold joint safety analysis and successfully localized the unsafe joints on the 3D point cloud data.