Development Of Small Magnetic Tile Mircro Defect Visual Inspection System

Magnetic tile materials are widely used in the development of science and technology in Industry 4.0,such as electric motors,generators and large transformers in the manufacture of electric power technology,as well as magnetic components such as electric ring plates,filters and oscilloscopes in power electronics.At present,most magnetic material manufacturers in China still use manual measurement and defect detection of magnetic materials.With the rapid development of industrial technology,the requirements of customer manufacturers for the quality and measurement accuracy of magnetic materials have gradually become higher.Traditional manual measurement and defect detection can not meet the requirements for the accuracy and speed of magnetic tile detection.In recent years,with the rapid development of machine vision,machine vision technology has gradually been put into industrial applications.Based on machine vision technology,this thesis studies the small magnetic tile size measurement and surface defect detection.Through a large number of offline and online experimental analysis,small magnetic tile micro defects have the characteristics of fast speed and good robustness.This thesis analyzes the current status of small magnetic tiles and the imaging characteristics of defects,and proposes three types of defect detection methods.According to the requirements of the enterprise,the relevant defect screening technical indicators were formulated,and the three types of small magnetic tile micro defect detection algorithms were designed as follows:Image preprocessing algorithm design:Because the initial image body part of the small magnetic tile obtained by the camera is not obvious,the gray level is uneven,and there are interference factors such as noise,this thesis passes the median filtering,mean filtering,frequency domain filtering and improved mean filtering.Denoising comparison is carried out.The mean squared error(MSE)and processing time ‘t'of the improved mean filtered image are smaller than those of other filtering methods?Therefore improved efficiency average filtering is selected to remove noise;After the image is segmented,the ROI ontology image part is extracted,and the denoised image is processed and compared by histogram method,iterative method and maximum inter-class variance method.The maximum inter-class variance methodwith the least time and the best efficiency is adopted.The ontology image part is obtained;in order to separate the background and foreground of the small magnetic tile,the effective area is extracted,and the Canny edge detection operator is used to compare the detection effect of the traditional edge operator with the edge detection effect of the Canny operator.The edge information extraction effect of the sub-magnetic tile is obviously superior to that of other traditional edge detection operators.Small magnetic tile defect detection algorithm design: For small magnetic tile imaging is not clear,detection difficulty,low contrast,complex texture background,uneven brightness,small defect area and many types of defects,a small magnetic tile micro defect detection method is proposed..Firstly,according to the influence of the curved surface of the small magnetic tile,the chamfering and the defect area on the imaging,the surface defects are determined by analyzing the difference of the grayscale,grayscale gradient and defect morphology of the defect area and the normal area in the image of the small magnetic tile surface.The types are divided into three categories;the first type of defects are mainly the missing size of the small magnetic tile,and the defect area can be directly extracted on the basis of threshold segmentation;the second type of defect is mainly that the gray area of the defect area is not much different from the normal area,and the defect The area is mostly linear,and it is easily interfered by the surface texture generated during the small magnetic tile grinding process.Texture suppression is performed first when detecting a second type of defect.The third type of defect is mainly that the gray level difference between the defect area and the normal area is relatively obvious,the edge area of the defect area is larger than the texture area,and the defects are mostly block-shaped.Secondly,according to the imaging characteristics of three kinds of surface defects,the defect morphological characteristics,and the relationship with the background area,the corresponding defect extraction methods are designed respectively.Finally,experiments are carried out using experimental devices in different external environments and defect types.analysis.The experimental results show that the micro-defective surface micro-defect extraction algorithm proposed in this study has good stability and robustness,and can accurately and quickly extract the defect area in the surface of small magnetic tile.The detection accuracy is 93.5%.Magnetic tile size measurement algorithm: The binary image of the magnetic tile is obtained by preprocessing.For the measurement of the length,width and thickness of themagnetic tile,the minimum circumscribed rectangle method and the moving point detection method are adopted.Due to the minimum circumscribed rectangle method,the measurement error of some magnetic tile edges with different bumps is too large;this thesis selects the dynamic point measurement method with better robustness and faster time.Finally,by fitting a multiple k and a deviation b to the ten sets of data of the measurement results by a least squares method,the measured value and the actual value form a linear relationship and return an actual value according to the measured value.