Over-height vehicle collisions with bridges continue to pose serious safety risks and impose significant infrastructure repair costs, despite the presence of clearance signage and record-based systems. Many current clearance datasets rely on legacy documentation or infrequent manual surveys, failing to reflect changes caused by resurfacing, structural modifications, or evolving roadway geometry. As a result, clearance information is often outdated, incomplete, or unreliable for operational decision-making.
This project develops a smart, AI-enhanced system that integrates mobile LiDAR scanning with machine learning models to automatically extract accurate bridge clearance and obstacle measurements. Neural network architectures will analyze 3D point cloud data to identify bridge geometry, detect vertical obstructions, and calculate clearances with sub-inch precision. The resulting data will be delivered through a user-friendly visualization and archiving platform designed for direct use by state Departments of Transportation. Pilot deployment in Massachusetts, in partnership with MassDOT, will demonstrate the system’s potential to improve clearance monitoring, reduce bridge strike risk, and modernize infrastructure safety workflows.