Optimizing Subpixel Vision-Based Displacement Measurement for Structural Health Monitoring Applications

Accurate displacement measurement is essential for structural health monitoring (SHM). With the rapid development of computer vision, video-based techniques have become attractive for non-contact displacement sensing. Among these approaches, subpixel displacement measurement algorithms can achieve high-precision even under low-resolution video conditions, thereby showing considerable potential for engineering applications. Although various subpixel displacement measurement algorithms have been developed, their performance differs considerably, highlighting the urgent need for systematic comparison and analysis. This study conducts a comparative analysis of four subpixel displacement measurement methods. First, the computational principles of the algorithms are systematically analyzed, and computer-simulated vibration videos are employed to evaluate measurement accuracy and computational efficiency under different video resolution conditions. Subsequently, laboratory experiments are performed to investigate the influence of measurement distance on accuracy. Finally, the robustness of each method against measurement noise is examined by comparing their accuracy under various signal-to-noise ratio conditions. The results demonstrate that subpixel algorithms can significantly enhance the accuracy of vision-based displacement measurement, with different methods exhibiting variations in efficiency, accuracy, and robustness. These findings provide new insights into the trade-offs between accuracy, efficiency, and noise tolerance, offering practical guidelines for selecting appropriate algorithms in vision-based structural displacement monitoring.