Research On Profile Detection System Of Medium Plates Based On Machine Vision

With the rapid development of China’s manufacturing industry,the medium medium plates has been widely used in various fields of manufacturing because of its excellent mechanical properties and high precision of finished products.At the same time,the plate shape quality requirements of medium plates products are also getting higher and higher.For medium plate profile detection,the traditional manual visual inspection already cannot satisfy the detection requirements of modern industry,therefore to seek a kind of based on machine vision detection system has become the development trend of flatness detection,also got a lot of attention and research institutions,and identity in iron and steel enterprises.In recent years,some machine vision detection technology has the advantages of non-contact,high detection precision,data processing speed,strong real-time performance and has been widely used in plate shape detection,under this background,this paper proposes a plate profile based on machine vision detection system used to achieve the measured plate profile detection.In this paper,a method of profile detection based on machine vision is presented.It consists of the incident angle of controllable line laser source,CMOS camera,magnetic grating measuring devices,through the acquisition of target image processing and then extract contour information,and then,in order to achieve speed data fitting of the plate profile in motion and display.This paper focuses on the plate speed acquisition,calibration,and calibration of the height of the camera image processing algorithms including bright line extraction algorithm research approximate extraction algorithm,bright line longitudinal position of the bright line endpoint position coordinates and sub-pixel extraction algorithm and accuracy analysis and so on.The details of this thesis include the following points:(1)The research status of machine vision detection technology and thick plate contour detection technology at home and abroad are introduced in detail.This paper discusses the background and significance of the research,finally through the analysis of the current research status in domestic plate detection and contour detection the expected requirements,proposed a plate profile based on machine vision detectionsystem,the system can realize laser source control,image acquisition,image processing,contour display and other functions.(2)According to the velocity of the plate,the plate speed measuring device of magnetic grid sensor based on the design and development based on LabVIEW speed acquisition interface,to deal with the speed of access,on the basis of no interference under the condition of speed and access interference under the access speed,the speed of data of the two cases,end coordinate data,fitting out the plate profile data.(3)In order to investigate the relationship between the height and the camera visual changes,through the design of a new camera calibration device and realizes the camera calibration calibration principle respectively,curve fitting of the calibration data using a function and two time function,by comparing the fitting parameters,get the fitting function is more accurate,so as to realize the the transformation between the image coordinates and plane coordinates.At the same time,the parameters of the camera are calibrated with the aperture imaging model.(4)According to the characteristics of plate contour detection system of the captured images,in order to accurately extract the image coordinates of the endpoints of the line,this paper presents an algorithm for vertical search,the algorithm can extract the bright line endpoint pixel coordinates and can calculate the number of end points.In order to further improve the detection accuracy,this paper studies an endpoint coordinate extraction algorithm based on two projections.(5)The contour data detected by the plate contour detection system,the detection accuracy is analyzed,including the extraction of the longitudinal coordinates of the bright line accuracy analysis and accuracy of endpoint coordinates of bright line extraction analysis,through calculation,longitudinal repeatability of bright line coordinates extraction can reach 0.2013 pixel repeat,detection of bright line endpoint extraction accuracy can reach 0.1689 pixel.(6)The research contents are summarized,and the future research work is putted forward.