Computer vision-based system of vibration monitoring for early-warning signals in dam breach

This study demonstrates both displacement and vibration monitoring of earth dams undergoing breach events using a computer vision-based measurement monitoring (CVMM) system. The system comprises a microcomputer equipped with computer vision algorithms to measure and compute changes in pixels of the chessboard images with 100 frames per second. In the study, the CVMM system was applied to indoor shake table and in-situ dam test using the analysis of vibration data from displacement, velocity, and acceleration. When dam breaching in the in-situ test, the maximum vertical displacement in the Y axis of the chessboard recorded by the CVMM system was 26.77 cm before the chessboard collapsed. There are two indicators evaluated for the dam breach in the case study. First is the amount of cumulative displacement that continued to increase until dam failure occurred. Second is the instantaneous energy of Hilbert–Huang Transform that showed a significant surge for an initial phase of dam breach. Consequently, the innovative CVMM system enables simultaneous monitoring of surface displacements, vibrations, and images. The vibration frequency of the CVMM instrument ranges from 0.1 to 50 Hz in the shake table and in-situ dam test, suggesting a non-contact monitoring and early-warning method in the dam breach test.