Output-only modal analysis of the Humen Bridge from video measurement

Video-based vibration testing is a promising non-contact method for structural health monitoring. This article develops an innovative methodology for estimating the structural modal parameters from videos without recovering the responses of structures. Firstly, an image processing (i.e., Gaussian smoothing) are adopted to smooth frames and alleviate the noises. Secondly, several regions of interest (ROIs) contained the projection area of structure are selected, and edge detection is performed to that to reduce the size of vision data. Then under some assumptions, the reduced vision data can be subjected to the frequency domain decomposition (FDD) similar to the structural modal analysis to identify the modal parameters of structure. Moreover, the structural mode shapes are recovering from such decomposition. The feasibility and accuracy of the proposed method were validated by a field experiment on the Humen Bridge in Guangzhou, China.