Research On Bolt Crack Detection System Based On Image Processing

The anchor rod that fixes the base of the wind turbine is an indispensable component in the wind power generation system and plays a crucial role in the development of the entire wind power generation system.Anchor rods are prone to small cracks during manufacturing and transportation.Currently,the detection of cracks mainly relies on manual vision,which has problems such as low detection efficiency,low accuracy,and high missed detection rate.This article uses a combination of magnetic powder and digital image processing methods based on the characteristics of anchor crack images for detection.The main research content includes:Based on the analysis of the current research status of magnetic particle testing technology and visual inspection technology based on image processing both domestically and internationally,and based on the characteristics of anchor rod cracks and the performance requirements for crack detection systems,an image processing based anchor rod crack detection system was designed,and the selection of software and hardware was completed.Extract the region of interest from the collected image of surface cracks on the anchor rod,set purple to 0 in the HSV color space,and solve the problem of noise interference caused by ultraviolet light irradiation when collecting anchor rod images.The weighted average method is used to grayize the image.Comparing the effectiveness of different filtering algorithms and improved filtering algorithms,objectively evaluating the filtering effect based on image quality indicators reflects the superiority of the improved algorithm,solves the problem of cumbersome adaptive median filtering algorithm and using noise points as output,and linearly changes the grayscale of the processed image,enhancing the contrast between cracks and background areas.Using a combination of Niblack local threshold and area threshold,the cracks in the surface image of the anchor rod are separated from the background information,and the morphological dilation operation is performed on the separated image to supplement the missing crack information due to image segmentation.Afterwards,by describing the geometric features of the crack information,cracks were screened out to achieve crack extraction from the surface image of the anchor rod.Based on the MATLAB software GUI platform,the software interface of the crack recognition system was written,and the software module design was completed.The collected bolt images were analyzed through detection experiments.The experimental results showed that the comprehensive detection accuracy of the bolt images reached 93.33%,and the missed detection rate was0%,meeting the design requirements.