Unmanned aerial vehicle (UAV) vision-based sensing has become an emerging technology for structural health monitoring (SHM) and post-disaster damage assessment of civil infrastructure. This article proposes a framework for monitoring structural displacement under earthquakes by reprojecting image points obtained courtesy of UAV-captured videos to the 3-D world space based on the world-to-image point correspondences. To identify optimal features in the UAV imagery, geo-reference targets with various patterns were installed on a test building specimen, which was then subjected to earthquake shaking. A feature point tracking-based algorithm for square checkerboard patterns and a Hough Transform-based algorithm for concentric circular patterns are developed to ensure reliable detection and tracking of image features. Photogrammetry techniques are applied to reconstruct the 3-D world points and extract structural displacements. The proposed methodology is validated by monitoring the displacements of a full-scale 6-story mass timber building during a series of shake table tests. Reasonable accuracy is achieved in that the overall root-mean-square errors of the tracking results are at the millimeter level compared to ground truth measurements from analog sensors. Insights on optimal features for monitoring structural dynamic response are discussed based on statistical analysis of the error characteristics for the various reference target patterns used to track the structural displacements.
This paper is the second of a set presented within the session: Findings from the CFS10 Multi-Hazard Test Program. The emphasis within this article is to highlight correlations between physical damage to nonstructural components and systems with measured response during a suite of 18 earthquake tests and 2 subsequent fire tests. This damage was documented within a 10-story cold-formed steel-framed building test specimen outfitted with various nonstructural components, including suspended ceilings, architectural finishes, a resilient stair system, windows and doors, pressurized fire sprinkler and gas piping systems, and roof-mounted mechanical equipment. The test building and multi-hazard protocol are described in the session companion paper. This paper provides test observations correlated with measured engineering demand parameters such as floor acceleration, building inter-story drift, or peak local temperature responses and offers related literature emerging from the project. Damage data and functionality checks will support development of fragility functions for these nonstructural systems for use in recovery-based frameworks.