TY - JOUR
T1 - A motion-blurred restoration method for surface damage detection of wind turbine blades
AU - Ying, Du
AU - Wu, Hongkun
AU - Garcia Cava, David
N1 - Funding Information:
This study has been supported by the Shenzhen Science and Technology Project, China (Grant No. JCYJ20190806153615091), and Xi'an University of Technology, China (Grant No. 102-451123002). The authors also want to thank Professor Shengxi Zhou for his financial support, and Haitao Xu for his help in the experiment.
Funding Information:
This study has been supported by the Shenzhen Science and Technology Project, China (Grant No. JCYJ20190806153615091 ), and Xi’an University of Technology, China (Grant No. 102-451123002 ). The authors also want to thank Professor Shengxi Zhou for his financial support, and Haitao Xu for his help in the experiment.
Publisher Copyright:
© 2023 Elsevier Ltd
PY - 2023/8
Y1 - 2023/8
N2 - In order to reduce expensive cost and downtimes, visual inspection method based on unmanned aerial vehicles has become a promising method for surface damage detection in wind turbine blades. Due to the fact that the linear velocity of points on the blade varies along its longitudinal dimension while the blade is in-operation, the captured blade image suers degradation caused by motionblur, and the degradation process of the image is nonlinear. To address this issue, an area-based deblurring method is rstly developed. The points on each divided region can be assumed to follow a uniform motion with the same linear velocity and blurring eect, which indicates that the nonlinear problem can be resolved by linear methods. Point spread function estimation is then conducted to each region to obtain the blur kernels, and image restoration of each region is carried out by using Wiener lter algorithm. Ultimately, image fusion is implemented to merge all the restored regions to realize blade image restoration. Experimental results and evaluation have demonstrated that the proposed areabased method can mitigate blurring effects of rotating blade images, thus further improve the detection accuracy for the on-line inspection.
AB - In order to reduce expensive cost and downtimes, visual inspection method based on unmanned aerial vehicles has become a promising method for surface damage detection in wind turbine blades. Due to the fact that the linear velocity of points on the blade varies along its longitudinal dimension while the blade is in-operation, the captured blade image suers degradation caused by motionblur, and the degradation process of the image is nonlinear. To address this issue, an area-based deblurring method is rstly developed. The points on each divided region can be assumed to follow a uniform motion with the same linear velocity and blurring eect, which indicates that the nonlinear problem can be resolved by linear methods. Point spread function estimation is then conducted to each region to obtain the blur kernels, and image restoration of each region is carried out by using Wiener lter algorithm. Ultimately, image fusion is implemented to merge all the restored regions to realize blade image restoration. Experimental results and evaluation have demonstrated that the proposed areabased method can mitigate blurring effects of rotating blade images, thus further improve the detection accuracy for the on-line inspection.
KW - Wind turbine blades
KW - Vision-based monitoring
KW - Motion-blur estimation
KW - Image restoration
KW - Surface damage detection
U2 - https://www.sciencedirect.com/science/article/pii/S026322412300595X
DO - https://www.sciencedirect.com/science/article/pii/S026322412300595X
M3 - Article
SN - 0263-2241
VL - 217
JO - Measurement
JF - Measurement
M1 - 113031
ER -