Non-Contact Monitoring of Rotating Machinery Using Optimized Camera-Based Stroboscopic Imaging

Abstract

Monitoring rotating machinery is essential for ensuring reliability and reducing maintenance costs. Cameras have recently gained attention as non-contact sensors due to their high spatial resolution and minimal interference. However, high-speed rotation introduces challenges such as motion blur, aliasing, and reliance on expensive highspeed cameras. This study proposes a camera-only stroboscopic imaging method for real-time, non-contact monitoring. By capturing frames at fixed angular intervals without external synchronization and applying optimized object tracking and phasebased optical flow (POF), the system estimates full-field, sub-pixel displacements. Phase-locked displacement data is used to identify rotor faults. Experimental validation on a lab-scale blade system confirms accurate displacement measurement and effective fault detection, offering a practical solution for aerospace, wind, and industrial applications.