EnhancedPVE: Video Phase-based Non-Identification Micro-vibration measurement for Bridge Cables

The video sensor-based method for bridge cable frequency measurement offers notable advantages in terms of speed, efficiency and non-contact advantages compared to the conventional acceleration sensor approach. However, accurately measuring the cable’s micro-vibration under natural excitation remains a challenge for the video sensor method. To address the problem, a novel phase-based frequency measurement method, named EnhancedPVE, is proposed, which leverages the signal characteristics difference between cable signal and noise signal in the temporal and spatial domains without the necessity of cable identification. Firstly, in the spatial domain, to preserve and enhance the cable phase while simultaneously attenuating the noise phase, a method combining bilateral filtering and phase-based motion amplification algorithm is designed based on the feature difference between cable edge phase and noise phase, which effectively enhances the discernibility of cable phase. Subsequently, in the time domain, to enhance the purity of the cable signal, the video image frame is partitioned into multiple subregions. Then, an innovative algorithm, named the maximum contribution combination algorithm, is devised to determine the contribution degree of each subregion to the cable signal, which differs from the traditional phase average weighting approach. Finally, the cable vibration obtained from all subregions are synthesized, forming the ultimate cable signal. The EnhancedPVE method is evaluated in laboratory and extensive outdoor experiments. Compared with the state-of-the-art methods, EnhancedPVE exhibits substantial improvements on the cable micro-vibration frequency measurement under natural excitation.