Research On The Detection System Of Enameled Wire Defects Based On Single Camera

Enameled wire is an important material in modern industry.Its surface quality will directly affect the comprehensive performance of the product.With the development of mathematics and computer,machine vision technology has also developed rapidly.Machine vision detection technology is applied to the detection of defects on the surface of enameled wire,instead of traditional manual detection and destructive contact detection,to achieve efficient on-line detection,so as to improve the output of enameled wire.According to the number,type and location of defects,it is of great significance to provide reference for the maintenance of production equipment.The core of machine vision detection system is digital image processing.The difficulty lies in panoramic image acquisition and high detection accuracy of defects.In this paper,a detection system of enameled wire defects based on single camera is designed.The main research contents are as follows.First of all,a single camera is designed to cooperate with four plane mirrors to realize360° panoramic acquisition of enameled wire images with a diameter of 2.5-4mm.According to the requirements of image acquisition,industrial cameras and lenses are reasonably selected,and double low angle circular light illumination is designed and proposed.Then,the image preprocessing algorithm of enameled wire is designed,the image is tilted by affine transformation,and extracted ROI to shield useless information.It is proved that power law transformation has the best enhancement effect on enameled wire image,and Gaussian filtering has the best denoising effect on image.Secondly,the threshold segmentation algorithm is used to segment the enameled wire area,the edge improved threshold segmentation algorithm is proposed to segment the defect area,analyze the shape characteristics of the defect area,propose the multi feature combination method to detect the classification defect,analyze the local gray level characteristics of the defect,and propose the gray level projection curve secondary fitting method to detect the defect,through the analysis and comparison,the two methods are combined to detect and classify the defects.Finally,C # and Halcon designed the detection software according to the idea of modularization,introduced the function and use method of the software human-computer interface,tested the system with enameled wire samples,the detection accuracy rate was97.7%,the error detection rate was 1.3%,the missed detection rate was 2.3%,when the detection speed was 1m/s,the experimental results show that the detection system meet the design requirements.