Natural frequency measurement of structures under ambient vibration using phase-based video motion analysis

Abstract

Structural vibration is commonly measured using accelerometers; however, depending on the structure, sensor installation may be physically difficult or impractical. In recent years, non-contact vibration measurement methods based on the analysis of video data captured by digital cameras have been proposed. In this study, a phase-based video motion analysis method is applied to estimate the natural frequencies of structures. Although this method is considered suitable for measuring small displacement amplitudes, the vibration response of structures under ambient excitation such as wind or traffic loads can be extremely small, which degrades measurement accuracy. To address this issue, long-duration video recordings are divided into multiple short-time segments, and video-based displacement analysis is performed for each segment. The displacement Fourier amplitude spectra obtained from individual segments are then averaged to achieve a more stable estimation of natural frequencies. The effectiveness of the proposed method is demonstrated through applications to a two-story frame model and a lighting pole.