Extracting Damage Indicator Features from Videos via Phase-Based Motion Estimation and Video Magnification
Recorded videos of the structures are considered as rich sources of information,which contain valuable data regarding the dynamic behavior of the test structure.However,the extraction of the desired parameters and variables from the videos often requires multiple layers of signal processing.Within this study phase-based motion estimation and video magnification are utilized to extract damage sensitive features for a wind turbine blade including natural frequencies and operational deflection shapes(ODS).Natural frequencies and the ODSs of the baseline and damaged wind turbine blade are estimated from the captured videos and it has been shown that these two pieces of information are sensitive to occurrence of damage such as change of mass.Moreover,the sensing methodology is non-contact that can make the testing procedure straightforward compared to other sensor systems such as accelerometers and strain gauges.The extracted features can be integrated with classical workflow of structural health monitoring(SHM)approaches such as machine learning to complete the loop of decision making regarding the presence of damages in the structure.
Phase-Based Video Magnification Computer Vision Wind Turbine Blade Damage sensitive Features
Aral Sarrafi Zhu Mao
Structural Dynamics and Acoustic Systems Laboratory Department of Mechanical Engineering,University of Massachusetts Lowell One University Avenue,Lowell,Massachusetts 01854
国际会议
The 7th World Conference on Structural Control and Monitoring(7WCSCM)(第七届结构控制与监测世界大会)
青岛
英文
1616-1619
2018-07-22（万方平台首次上网日期，不代表论文的发表时间）