Automatic checks from 3D point cloud data for safety regulation compliance for scaffold work platforms

Abstract

Abstract Fall from scaffolds is one of the leading causes for injuries and fatalities in the construction industry. To prevent fall from scaffolds, toe-boards and guard-rails must be installed on scaffold work platforms according to relevant safety regulations. Traditionally, the checking of safety regulation conformity relies on manual observation, which is inefficient and inaccurate. To address the limitations of manual checking, this study proposes a technique to automatically check whether scaffold work platforms conform to the safety regulations based on 3D point cloud data. The proposed technique first detects the location of scaffolds from the point cloud data by finding the vertical scaffold components, known as uprights. Then, scaffold work platforms, which are planar and horizontal components, are detected based on the histogram of the Z values of the point cloud data. Once the work platforms are extracted, the toe-boards and guard-rails are detected along all the four sides of each work platform. Then, the detected toe-boards and guard-rails are checked to identify any violation of safety regulations. Validation experiments were conducted on a point cloud dataset acquired from a construction site in Singapore. The experimental results show that the proposed technique could successfully detect scaffolds and work platforms from point cloud data, and extract toe-boards and guard-rails for safety regulation checking. All the violations of safety regulations in the point cloud data were successfully identified using the proposed technique.