Phase-Based Vibration Frequency Measurement for Civil Structures with Multidimensional Camera Motion Suppression

Phase-based vibration estimation using video has attracted significant attention in civil engineering due to its noncontact characteristic and high sensitivity. However, camera motion often introduces spurious phase disturbances, significantly reducing the accuracy of vibration frequency estimation. To address this issue, a novel phase-based vibration frequency measurement with multidimensional camera motion suppression (PVMC) approach is proposed, which integrates directional, spatial, and temporal dimensions to effectively suppress camera motion interference. Specifically, in the directional dimension, a complex steerable filter is used to extract phase responses along the structural vibration direction, thereby suppressing motion interference from nontarget directions. In the spatial dimension, a region-growing algorithm is employed to segment structural vibration target regions and remove background interference caused by camera motion. In the temporal dimension, based on the periodic characteristics of vibration signals, a sparse coding method leveraging the fast iterative shrinkage thresholding algorithm (FISTA) is designed to suppress camera motion interference and accurately extract structural vibration frequencies. Experimental results from both indoor and outdoor scenarios demonstrate the effectiveness of the proposed method. It outperforms existing state-of-the-art techniques in vibration frequency measurement accuracy under camera motion interference, demonstrating its suitability for engineering applications.