Image recognition method for blade fault of large offshore wind turbine
Aiming at the problem of lack of a large number of actual fault image training samples during the fault diagnosis and modeling of offshore wind turbine blades,an image recognition method for offshore wind turbine blade faults based on small data sets is proposed.In this method,the K-means clustering algorithm is improved to identify blade segmentation according to the color and shape characteristics of blades and their faults in wind turbine blade images,an adaptive algorithm is designed to adjust the Canny operator parameters to identify the segmentation of early fault areas on the blade surface,and the K-means clustering algorithm is used to extract the color and shape features of faults and design corresponding classifiers to achieve fault classification.Simulation examples show that this method is effective for the identification of early faults on the blade surface,and can provide an accurate diagnostic model for the blade fault identification of offshore wind turbines on the basis of a small number of fault samples,which can improve the operation and maintenance efficiency of offshore wind farms.
offshore wind turbineblade faultimage recognitionsmall data set
万方数据
维普
