Automated Underground Utility Mapping and Compliance Checking Using NLP-Aided Spatial Reasoning

The recurrent underground utility incidents (e.g., utility strikes and utility conflicts) highlight two underlying causes: failure to comply with spatial rules prescribed in utility specifications and unawareness of utility locations. It is critical to address these two causes to prevent utility incidents. This paper presents a framework that integrates natural language processing (NLP) and spatial reasoning to infer the vertical positions of underground utilities from textual utility specifications and plans, and to automate the utility compliance checking. The natural language processing (NLP) algorithm extracts the spatial rules specified in textual utility documents, and converts the extracted spatial rules to a computer-interpretable format. The spatial reasoning scheme models the spatial components in a spatial rule as topological, directional, and proximate relations, and executes the extracted spatial rules in a geospatial information system (GIS) to automate the depth estimation and the compliance checking. Several examples are presented to prove the concepts.