Machine learning (ML) and artificial intelligence (AI) have significantly impacted numerous fields through their ability to tackle challenges with remarkable computational efficiency. In bridge engineering, ML/AI techniques have been employed to enhance the efficiency of the structural design phase, aid in the selection of optimal bridge types, produce cost estimates, conduct real-time structural health monitoring, predict structural response and deterioration, reconstruct data for comprehensive health assessment, and prioritize maintenance efforts. This research project applied ML/AI techniques to automate the process of extracting data and features from drawings, tables, and text blocks contained in bridge plan sets using state-of-the-art computational algorithms. The research was motivated by the critical need to report bridge inventory information to the Federal Highway Administration (FHWA) in compliance with National Bridge Inspection Standards (NBIS) reporting requirements. This research project produced a novel platform that automates the process of reviewing bridge plans to identify, extract, and report select engineering details. While the automated extraction of details from engineering documents can be a complicated task for machines due to the complex nature of plan sets, a combination of several deep learning models and various image processing techniques provided a platform to successfully extract details of interest. Furthermore, using the general models and functions developed in this research, the platform can be customized for different transportation agencies, following their formats and practices to capture bridge details available in their plan sets. 

The Final Report is available here.