Condition monitoring (CM) of rotating machinery is critical for maintaining operational efficiency, reducing maintenance costs, and preventing unexpected failures. Traditionally, CM has relied primarily on vibration-based analysis using contact sensors. While effective, such approaches require physical installation and may interfere with system operation. Recent advances in vision-based techniques offer significant advantages, including non-contact measurement, full-field analysis, and high spatial resolution. This study presents a novel stroboscopic imaging-based method for monitoring the condition of rotating blades. Unlike conventional stroboscopic approaches that depend on costly lighting systems and high-speed cameras, the proposed technique operates with low-cost imaging hardware by combining phase-synchronized acquisition with vision-based measurement. The system employs a vision sensor triggered by phase-locked signals synchronized with the rotational speed to produce stroboscopic video sequences. These sequences are analyzed using a phase-based optical flow algorithm to estimate micro-displacements of rotating structures marked with designed markers. The resulting displacement fields are processed to extract diagnostic features, which are then used to classify structural conditions as normal or damaged. Experimental validation demonstrates that the method can reliably detect critical fault modes, including bolt loosening, mass imbalance, and shaft misalignment. The results highlight the potential of this vision-based approach to deliver robust, accurate, and efficient CM of rotating machinery, offering clear advantages over traditional contact-sensor methods. The proposed technique advances the development of non-invasive CM solutions and can be extended to a broad range of rotating mechanical systems.
Jeong et al. (Thu,) studied this question.