Research on Key Technologies of wind turbine blade health monitoring and maintenance, a key project of science and technology plan of Inner Mongolia Autonomous Region in 2020
Aiming at the early damage characteristics such as wear on the surface of wind turbine blades and the problems of high cost and low efficiency in traditional damage detection methods, a wind turbine blade damage detection system based on the combination of machine vision and image processing is designed. The machine vision experimental platform is built to complete the image acquisition and processing of the damaged blade of the wind turbine. The HSV is used for color plane extraction, convolution operation, highlight operation and filtering. The minimum uniformity measurement method in the automatic threshold segmentation method is selected for threshold segmentation. Finally, the mathematical morphology denoising is used to eliminate corrosion, expansion Open operation and other operations to complete feature extraction. An intelligent image recognition system of wind turbine blade based on LabVIEW is designed. Through the debugging of damage feature recognition effect after image processing, the performance test is completed. The experimental results show that the image processed based on the algorithm has an accurate recognition rate of 92.3% in the designed recognition system, and the actual length of crack damage is measured, and the maximum absolute error is 3mm. The system meets the requirements of blade damage detection, realizes the image processing and identification of wind turbine blade surface cracks, contour wear and other damage, marks, counts and measures the damage, and realizes non-destructive flaw detection, which provides method reference and technical support for image processing and system design for megawatt wind turbine blade damage detection.
王一博,韩巧丽,张曦文,等. 基于机器视觉的风力机叶片损伤检测系统[J]. 科学技术与工程, 2022, 22(12): 4879-4886.
Wang Yibo, Han Qiaoli, Zhang Xiwen, et al. The blade of wind turbine based on machine vision[J]. Science Technology and Engineering,2022,22(12):4879-4886.