ABSTRACT
Bridge inspections of medium or large scale are often cumbersome, expensive, and time-consuming. With a large number of bridges in the United States that require at least a bi-yearly inspection, there is a need to improve bridge inspection techniques to save time and reduce costs. Unmanned Aerial Vehicles (UAVs) have made tremendous advancements in recent years to allow for better data collection with enhanced sensors, more controllability with precise global-positioning-systems and inertial measurement units, and increased safety with omnidirectional sensors to avoid collisions while becoming more affordable. In current bridge inspection practice, UAVs have been used as “eyes-in-the-sky,” simply assisting inspectors to view bridges or other structures from different vantage points with the inspectors still taking measurements and making decisions with the traditional techniques. However, to take full advantage of the UAV’s capabilities and allow for the UAV to perform and quantitate inspections automatically to create a more streamlined workflow, there is a need for more robust data processing of the information attained by the UAV. A streamlined decision-making support framework is proposed that uniquely integrates UAV-based field inspection, automated damage identification, and establishment of an element-wise As-Built Bridge Information Model (AB-BrIM) for the damage documentation. In this framework, a UAV platform with optical sensors first collects the data. Next, an automated damage detection algorithm that highlights cracks and spalling is developed to quickly extract quantitative information (i.e. type, size, amount, and location). Finally, a 3-D point cloud is created with photogrammetry and then segmented into identified structural elements (e.g. beam, girders, deck, etc.) to serve as a base for the AB-BrIM. The identified damage information is automatically linked to each element. The resulting AB-BrIM with 3-D visualization of element-wise, quantitative damage information offers a transparent condition evaluation and thus can greatly ease the planning of repair/maintenance.
