Automated procedure for extracting safety regulatory information using natural language processing techniques and ontology

Traditionally, the identification of design-related hazards inherent in design drawings has been performed manually by safety experts. However, this manual approach may lead to incomplete, inaccurate, or incompatible results because of its repetitive, time-consuming, and error-prone process. For this reason, automating the safety design review process is expected to save time and reduce human interpretation errors. In this paper, we address this issue by formulating a procedure of ontology-based information extraction using natural language processing (NLP) techniques and apply it to safety review in the design phase. Specifically, construction safety requirements are identified from textual regulatory documents, and then, are converted to machine-readable format. The proposed approach was applied to extract hazard information from two different types of regulatory documents. Preliminary results demonstrate that this approach is effective in automating the hazard information extraction without the manual interpretation from safety experts.