Bridges, Tunnels and other Transportation Structures
Full Session with Abstracts
Bridges should be inspected on a regular basis in order to quantify and track damage, and create effective rehabilitation and replacement plans. With continued improvement in the areas of mobile computing and sensing technologies, autonomous 3d mapping vehicles (both ground & aerial) are increasingly being recognized as a viable method for quickly collecting the visual information required for structural inspection. These autonomous sensing platforms would result in faster bridge inspections that produce more reliable data, ultimately leading to lower costs and better knowledge on the current state of managed infrastructure. However, before the paradigm shift from manual on-site inspection to off-site review using autonomously collected data can occur, it must be established that the quality of inspections conducted using autonomously collected data are at least at parity with those conducted by an on-site inspector. In this paper, we address this gap by presenting a detailed comparison of the accuracy and precision of a bridge inspection completed using only data collected by an autonomous ground vehicle with a traditional bridge inspection completed by an on-site inspector. In this study, the platform used for autonomous data collection uses multiple 3d lidars, an optical camera, and a thermal camera to create a detailed map of the bridge. Defects are subsequently identified, classified, and quantified using a semi-autonomous procedure. The traditional on-site bridge inspection used for comparison was obtained from publicly available reports. Preliminary results indicate that the majority of the defects identified by an on-site inspector are easily identified and quantified using the data collected by the mobile sensing platform.