Structural frequency identification based on broad-band phase-based motion magnification and computer vision

In order to solve the problem that the existing visual technology is difficult to accurately measure the small vibration of the structure, this paper proposed a method combining the broad-band phase-based motion magnification (BPMM) and sub-pixel template matching for structure vibration measurement and frequency identification. Firstly, the vibration video of the structure was captured by a camera with initial image noise removal. Secondly, the broad-band filtering was executed in a wide band containing all the frequencies of interest. A suitable magnification factor was selected for the magnification of the small vibrations. The displacement time-history of the structure was then extracted from the amplified video using the sub-pixel template matching algorithm. Finally, the extracted vibration displacement was normalized to obtain the actual displacement time-history, and the vibration frequencies of the structure were obtained using the fast Fourier transform (FFT). The results show that the BPMM algorithm can effectively remove the noise in the vibration video, which improves the signal to noise ratio of the image motion. The vibration frequency of the structure can be accurately identified from the video image after motion amplification. The proposed method in this paper was verified by the actual measurement results of a three-story frame structure in laboratory and a field bridge, and the identification errors are within 1% and 5%, respectively. The proposed method can achieve the blind magnification of small vibrations without the priori information of structural frequency and has the advantages of good noise robustness, low computational cost, and strong applicability. The research results can provide new choice for the measurement of small vibrations of structures. © 2023 Editorial Office of China Civil Engineering Journal. All rights reserved.