Defect?oriented supportive bridge inspection system featuring building information modeling and augmented reality

摘要

Bridges are indispensable links of transportation infrastructure systems, and inspections play a critical role in maintaining bridge components in the state of good repair. Through a survey of bridge inspectors, the authors revealed that visual inspection techniques are the prominent inspection method but result in inaccuracy and ambiguity due to high variances among inspection results; modern inspections using drones and robots could improve efficiency but pose new challenges and do not reduce subjectivity. As a result, a novel, building information modeling- and augmented reality-based supportive inspection system (BASIS) that objectively captures bridge defects is proposed and validated. On-site inspectors can access the bridge model containing historical defect information (defect type, length/width/depth, and location) and overlay relevant content on the actual infrastructure through BASIS for inspection data collection with more accuracy and less ambiguity. A proof-of-concept prototype of the BASIS for bridges was developed as an android application and verified by bridge inspectors for effectiveness on a small pedestrian bridge. It was found that BASIS was able to collect accurate inspection data irrespective of the level of experience of the user, thusly minimizing the data subjectivity caused by differences among inspectors' judgment and/or human errors. This research explores the utilization of emerging tools to collect bridge condition information in a more comprehensive and objective manner. Collected information can be further integrated it into a digital model that reflects the bridge's most accurate and up-to-date condition, heading toward a digital twin of the physical infrastructure. The proposed system may also be adapted for other types of infrastructure (e.g., dams, levees, and railroads) that also require routine inspections.