Bridges, Tunnels and other Transportation Structures

Full Session with Abstracts

505500 - Fatigue crack sensing of steel structures using UAV-based computer vision

Friday, April 26
3:30 PM - 5:00 PM
Location: Bayhill 29-30

Fatigue cracks are one of the major concerns of structures that are subject to repetitive loadings. If not appropriately monitored, these cracks can continuously grow to a critical level and jeopardizes the structural integrity. Manual inspection of fatigue cracks is challenging due to the small size of crack openings and the large scale of civil structures. Over the past decades, many sensor-based structural health monitoring (SHM) and nondestructive testing (NDT) techniques have been proposed to achieve more reliable and autonomous fatigue crack detection. However, extensive human involvement for sensor installation and operation may still be needed. In this paper, we propose a non-contact autonomous fatigue crack monitoring strategy based on computer vision and unmanned aerial vehicles (UAV). To detect fatigue cracks using computer vision, a short video of a region of interest (ROI) of the structural surface is recorded. Then, feature tracking algorithms are applied to detect and track feature points of each video frame, enabling precise tracking of the monitored structural surface motion under fatigue loading. Subsequently, the tracked surface motion is analyzed to reveal the differential movement pattern caused by the existence of fatigue cracks. Through motion surface tracking and analysis, the location and size of fatigue cracks can be obtained. To enable the integration of the proposed computer vision-based fatigue crack detection method with autonomous platforms such as UAVs, the additional motion introduced to the video due to the nonstationary movement of the UAV needs to be compensated. To this end, geometric transformation is applied to align the feature points of each video frame to the same coordinate system, such that the differential motion pattern due to fatigue cracks can be reliably extracted. The proposed strategy is validated using a compact, C(T), specimen inspected by a UAV in a laboratory setup.

Jian Li

Assistant Professor
University of Kansas

Jian Li is an assistant professor in the Department of Civil, Environmental and Architectural Engineering at the University of Kansas. He received his MS in 2007 and BS in 2005 from Harbin Institute of Technology in China and his Ph.D. in 2013 from the University of Illinois in Urbana-Champaign, all in Civil Engineering. His research focuses on both theoretical and experimental developments of advanced sensing and health monitoring techniques to improve the resiliency and sustainability of civil infrastructure under operational and extreme loading conditions. His specific research interests include vibration-based damage detection and model updating, wireless smart sensor networks, innovative sensing techniques, computer vision, uncertainty quantification, risk assessment and mitigation. His research is currently funded by various agencies including National Cooperative Highway Research Program, Transportation Pooled Fund Program, Mid-American Transportation Center, and Kansas Department of Transportation.

Presentation(s):

Send Email for Jian Li


Assets

505500 - Fatigue crack sensing of steel structures using UAV-based computer vision



Attendees who have favorited this

Please enter your access key

The asset you are trying to access is locked. Please enter your access key to unlock.

Send Email for Fatigue crack sensing of steel structures using UAV-based computer vision