TY  - EJOU
AU  - Hou, Qingmin 
AU  - Xiao, Guanghua 
AU  - Liu, Qiang 
AU  - Su, Chunhui 
AU  - Yan, Ting 
AU  - Sun, Yafen 
AU  - Zhang, Peng 

TI  - A Novel Smart Vision-Based Sensing System for Structural Health Monitoring: Robust Angle Measurement of Offshore Rotating Machinery with Data-Driven Calibration
T2  - Structural Durability \& Health Monitoring

PY  - 
VL  - 
IS  - 
SN  - 1930-2991

AB  - To address the critical need for reliable condition monitoring of rotating components (e.g., crane slewing bearings, thrusters) in harsh marine environments characterized by strong electromagnetic interference, salt spray, and persistent vibration, this paper presents a novel smart sensing system for high-precision rotation angle measurement. The proposed intelligent inspection system employs a vision-based principle: a laser spot, projected from a turntable synchronized with the measured shaft, forms a circular trajectory on a fixed Complementary Metal-Oxide-Semiconductor (CMOS) image sensor. A robust machine vision algorithm, integrating an improved Canny-Kirsch operator and a least-squares ellipse fitting method, achieves sub-pixel spot center localization under dynamic conditions. To overcome systematic nonlinearities, a data-driven polynomial compensation modelis developed, demonstrating a data-driven calibration approach for enhanced accuracy. Experimental results demonstrate that after compensation, the system achieves a measurement accuracy of 0.9°, with sensitivity, resolution, repeatability error, and hysteresis of 4.25 Counts/°, 0.0017°, 1.83%, and 1.09%, respectively. While matching the precision of traditional encoders, the system exhibits superior immunity to electromagnetic interference and environmental robustness. This work, therefore, presents a practical and smart sensing solution with potential for the structural health monitoring and predictive maintenance of marine infrastructure, contributing to the development of robust sensor systems for industrial applications.
KW  - Novel smart sensing; structural health monitoring (SHM); vision-based sensor; offshore platform; data-driven modeling; nonlinear compensation

DO  - 10.32604/sdhm.2026.084518