Research On The Visual Detection Of Incomplete Ceramic Film Based On Ultrasonic And Machine Vision

Ceramic membrane is widely used in the pretreatment process of reverse osmosis desalination because of its excellent chemical stability and anti microbial properties.However,in the process of its preparation and use,it is often accompanied by surface defects and internal defects-incomplete ceramic membrane,which seriously affects its own mechanical strength and material mechanical properties,thus reducing the desalination efficiency and the service life of ceramic membrane.Therefore,how to realize the automatic detection of ceramic film product quality has become an urgent problem to be solved.In this paper,the non-destructive testing of the incomplete ceramic film is taken as the goal.Based on the analysis of the research status of the fusion ultrasonic and machine vision testing technology,the system design of the non-destructive testing of the incomplete ceramic film is realized.In order to meet the requirements of non-destructive testing of incomplete ceramic film,on the basis of the research on the surface and internal defect detection technology of ceramic film at the present stage,combined with the characteristics of machine vision and ultrasonic testing technology,this paper designs a hardware system of incomplete ceramic film detection combining ultrasonic and machine vision.In the system,einscan-se scanner and phascan phased array platform are used to obtain the surface image and internal data of the incomplete ceramic film.The surface image can be used for the classification and characterization of the surface defects of the incomplete ceramic film based on machine vision.The internal data can be used for the 3D reconstruction and visual analysis of the internal defects of the incomplete ceramic film based on ultrasonic detection.Aiming at the classification and characterization of the surface defects of incomplete ceramic films,this paper designs a classification method based on mixed features:the obtained surface images of ceramic films are preprocessed by images,and the eigenvalues of the defects of ceramic film samples are calculated based on the six features of area,aspect ratio,perimeter,roundness,diagonal length and second-order moment of angle,and the defect characteristics are established There are four kinds of defects:scratch,crack,slag and pit.Finally,the classification model of ceramic film defects is established by BP neural network classifier.Based on 80 groups of sample data,the average accuracy of classification for ceramic film surface defects is 90.08%.In order to realize the 3D visualization display and evaluation of the internal defects of the incomplete ceramic film,this paper proposes a 3D reconstruction method of the internal defects of the ceramic film based on the fusion triangle matrix synthetic aperture focusing algorithm.The ultrasonic image of the ceramic film based on the fusion triangle matrix synthetic aperture focusing algorithm is obtained,and the acquired B-scan image is preprocessed and the 3D reconstruction based on the defect contour can be completed Three dimensional visualization of defects in ceramic films.This algorithm can effectively improve the detection efficiency by reducing nearly half of the amount of data acquisition and calculation,while retaining the characteristics of high imaging resolution of synthetic aperture focusing algorithm.The experimental results show that the error range of diameter measurement of inner hole type defects is between 2.0%and 3.0%,and the error range of diameter and length measurement of crack defects is between 2.0%and 4.0%.