Structural identification of bridges using computer vision techniques

Currently, most of the studies about computer vision-based Structural Health Monitoring (SHM) focus on utilizing the visual tracking techniques to monitor the responses (structural output) of structures under various excitation, such as vibration, displacement/deflection and strain. By employing SHM data, Structural Identification (St-Id) can be carried out to identify the intrinsic characteristics of structures, such as health condition, load capacity, damages, etc. Ideally, St-Id utilizes not only the responses (structural output) of structures, but also the external loads (structural input). Vehicle loads are the most common external loads for bridge structures and it is essential to know the vehicle load distribution for a complete St-Id. In this work, a vision-based structural identification framework for bridges is proposed which combining the vision-based displacement measurement and vision-based vehicle load localization. Unit Influence Line (UIL) is extracted as an indicator of structural identification and the proposed methods is validated on a footbridge on the campus of University of Central Florida (UCF).