Research On The Defecting Surface Detections Of Magnetic Ring Based On Machine Vision

China is a manufacturing country of magnetic material,it have many material production enterprises.At present,most product material enterprises are still using the traditional manual detection to measure and defect of magnetic.As China's manufacturing industry to upgrade,enterprise are required the products of higher quality,precision and reliability,the traditional magnetic products based on artificial measurement and defect detection has been more and more unable to meet the accuracy and speed of detection.Although ultrasonic,eddy current and other non-destructive techniques have been applied to the product defect detection,but its high cost,and slow speed of detection.With the development of science and technology,it is possible use machine vision to measure the product,in addition to use of machine vision have fast speed and high accuracy to ring size measurement and defect measurement.Based on the background of the development,This paper were studied by the machine vision to detect the magnetic measuring and surface defects,and the related equipment of core size measurement and surface defect detection are developed.First,this paper analyzes the research status of magnetic surface defect detection and put forward the key technical points of this topic.According to the manufacturers and research group of the requirements are constitute the relevant technical indicators,and design of the magnetic defect detection algorithm according to the requirements,as shown below:Design of the image preprocessing algorithm:through the camera to obtain the magnetic image acquisition,there are uneven gray level,noise and other interference factors.Therefore,the image preprocessing must be carried out before the test to reduce the influence of interference factors.First,Using the magnetic image gray processing piecewise linear gray level transformation.Then,use this improved median filter implementation of the noise reduction of magnetic image.Finnlly,in order to separate the magnetic background and extracting the effective area,this paper introduced threshold segmentation and edge detection based on the method of OTSU,using this method to segmentate the magnetic images.The design of ring size measurement algorithm:through the pretreatment of the image of the magnetic ring obtained the segmentate two value image,By respectively using the Hough circle detection,three point method and least square method respectively on the internal and external contours obtained by fitting the magnetic ring inner and outer diameter.Finally through thecomparison of the measurement results and the running time of the algorithm,choose the least squares fitting ring ring outer diameter of inside and outside contour measurement.This paper through the magnetic ring side pixel area obtained the size of the magnetic ring side height.The design of magnetic surface defect detection algorithm: According to the magnetic surface image with complicated texture and defect types,features uneven,this paper puts forward two kinds of magnetic surface defect detection algorithms are as follows:(1)Through the analysis of gray difference defects and normal region magnetic surface in the image will be divided into two types of magnetic surface defects;The first category is the difference between the gray value and the normal area,which is easy to segment.The second type is the difference between the gray value of the defect and the gray value of the normal area.For the first type of magnetic ring defect,First of all,using OTSU implementation of image segmentation threshold ring.Then according to the measurement of the magnetic ring outer diameter and center ring structure mask image.Finally,the defect is extracted by mask image operation.For the second types of magnetic ring defects,First,image segmentation and operation with the original image using OTSU ring,shielding ring background area.Then use the Canny edge detection operator to extract the edge information of magnetic surface defects.The interference information mask image operation shielding ring surface texture and magnetic contour edge formation,extraction of magnetic surface defects.Last paper,In terms of different illumination,specification,defect type,a large number of online experiments are carried out using the developed prototype.The experimental results show that the defect extraction algorithm has good stability,strong robustness,can quickly and accurately extract the defects of the regional magnetic surface image,the accuracy of surface defect detection was 95.3%.(2)this paper presents a method for extracting magnetic surface defects and improved adaptive Canny algorithm based on mask.First of all,based on the analysis of magnetic surface image in different regions of gray and gradient features on the magnetic structure by fitting the contour mask image,so that interference shielding ring background area;Secondly,an adaptive Canny edge 8 neighborhood anisotropic filtering improved detection algorithm based on interference suppression of magnetic surface texture;Then,using digital image enhancement edge connected domain morphology,and use the extracted magnetic surface defect mask image structure;Finally,a large number of online experiments are carried out by using the developed prototype.Experimental results show that,this paper defects extraction algorithm has good stability and strong robustness,can quickly and accurately extract the defects of the regional magnetic surface image,the accuracy of surface defect detection was 97.3%.