Research On Flange Surface Defect Detection Technology Based On Machine Vision

With the mass production of mechanical parts,the production of flanges in parts is also increasing.The traditional manual flange surface defect detection method can not meet the needs of modern industrial production,so the machine vision technology is used to detect the surface defects of the flange.The application of this technology can greatly reduce the use of human resources,and has unique advantages such as high efficiency,high precision and non-contact detection.At present,the development of machine vision technology in China is changing rapidly and the momentum is rapid.However,under complex conditions,especially when the contrast between the workpiece and the environment is poor,there are still many urgent problems to be solved in the application of machine vision technology.In this paper,a set of flange surface defect detection system is designed,including two parts of machine vision detection platform and defect detection software.It can detect the four types of defects on the surface of the flange,such as point,line,surface and defect.It realizes the classification and accurate positioning of defects,and has good practical value.(1)In order to improve the contrast between defects and flange surface,LED ring light source and low angle illumination are used.This scheme can improve the accuracy of flange surface defect detection and reduce the probability of false detection and missed detection in the process of defect detection.At the same time,the scheme can also avoid the shadow of the flange surface,and can better identify the edge information of the flange surface.(2)Through four kinds of image quality evaluation indexes,the objective evaluation criteria of image change before and after image processing are obtained.Through the histogram analysis of the normal flange surface,the histogram characteristics of the background image of flange defects are obtained.By comparing the four filtering algorithms,the median filtering has obvious effect in the experiment of flange surface defect detection.By comparing five edge detection algorithms,Roberts edge detection has the best experimental effect.By calculating the characteristic parameters of the defect region,four different types of defects can be distinguished by analyzing the length-width ratio,area and defect location of the defect region.The location and type of defects can be directly obtained by using the defect location algorithm to accurately locate the defect area.(3)The human-computer interaction interface of the detection system was designed,and the algorithm involved in this paper was verified by using the flange surface defect detection experimental platform.Through the accuracy calibration of the flange surface defect detection system,the ratio coefficient between a single pixel and the actual size is determined to be 0.03068 mm/px.Through the flange surface defect detection experiment,the system defect detection measurement accuracy can reach 0.14 mm.At the same time,the defect detection platform can also display the coordinate information of the defect area and export the parameters of the current defect in the human-computer interaction interface.