Research On Surface Defect Detection Technology Of Wind Turbine Blade Based On Image Processing

Wind energy is a green and environmental renewable energy,which plays an important role in optimizing China's energy structure.Wind turbine blades are key components for capturing wind energy.Due to the long-term exposure to the harsh open air working environment,the surface will inevitably produce various defects,which will lead to serious consequences if not repaired in time.At present,domestic wind farms generally adopt the methods of telescope observation and visual inspection of high-altitude hanging inspectors.They only rely on experience to judge whether there are abnormal conditions.This mode has the problems such as poor safety,low efficiency,limited observation angle and difficult to find early defects.Moreover,the inspection cycle is long,and the equipment may still fail during the equipment maintenance interval.In recent years,unmanned aerial vehicles(UAVs)have been used to inspect wind turbine blades in some wind farms,but they still remain in the way of manually operating UAV for real-time observation or manual detection after image acquisition.They have not realized the automatic detection of wind turbine blade defects.It still requires a lot of manpower and time costs,and can not effectively shorten the inspection cycle.In this context,this paper develops a wind turbine blade surface defect detection system based on image processing technology.The main research contents are as follows.Aiming at the motion blur image generated during the process of image acquisition by UAV,a blind image deblurring algorithm based on dark channel and low rank prior was proposed.The algorithm firstly used the sparseness of the dark channel of the image to estimate the intermediate latent image.And the weighted low rank prior constraint was introduced to suppress the noise in the intermediate latent image,so as to improve the accuracy of the fuzzy kernel estimation.Then,the accurate blurring kernel was obtained by the optimization strategy of alternating iterations.Finally,the Hyper-Laplacian prior method was used to get the clear image.The experimental result shows that the proposed algorithm can effectively suppress the noise and ringing artifact and has a good restoration effect for a motion blurred image.The characteristics of different types of defects on the surface of wind turbine blades were studied and a classifier was designed.Firstly,through the edge detection,neighborhood search and morphological operations,the defect part in the image was extracted;then the shape factor of the defect contour,the ratio of the long and short diameter,the gray mean and the gray variance were calculated,and the defects were classified by the difference of the characteristics of different defects.The experimental results show that the method can classify the surface defects of wind power blades accurately.Aiming at the problem that it was difficult to determine the defect position by segmented image acquisition,a method of surface defect location of wind power blade combined with image stitching and coordinate transformation was proposed.The method combined the images acquired by the segmentation into an image of complete blade,calculated the position of the defect in the complete blade image,and obtained the actual position of the defect through coordinate transformation by using the displacement data when the image was acquired by UAV.And,in order to improve the performance of image mosaic,a sequence image overlap region estimation algorithm was proposed.The experimental results show that the method can accurately calculate the position of the defect on the wind turbine blade.Using Qt Creator as the development environment and Halcon 13.0 as the assistant development tool,the software of defect detection system was developed by C++ programming language.