Natural language-extracted and BIM-referenced knowledge base for construction quality inspection via augmented reality

Construction quality is of upmost importance for delivering well-performed civil structures. Inspection offers a critical means to ensure construction quality. However, its effective implementation relies on an excess of domain knowledge that usually takes years to accumulate, making inspection expensive and challenging to conduct. This paper introduces a construction quality inspection knowledge base that empowers inspectors with easily accessible and intuitive construction requirement information. Natural language processing (NLP) is applied to automatically extract knowledge from construction documents such as specification and regulatory files. The extracted knowledge is linked to the building information model (BIM) using proposed association methods and semantic similarity matching. The natural language-extracted and BIM-referenced knowledge base (NLBIM-KB) is integrated into an augmented reality (AR) interface, which provides a freehand tool to assist inspectors' decision-making via on-demand construction knowledge extraction.