Research On Geometrical Inspection Of Mechanical Parts Based On Machine Vision

In recent years,machine vision technology has gradually become a hot research field of many researches.With the advantages of high detection accuracy,high efficiency,and non-contact,machine vision technology has been widely used in the detection of the appearance and defect of mechanical parts.The quality inspection of parts is an important part that must be carried out before the parts are used.Only the qualified parts can meet the requirements for use.The traditional detection methods cannot meet the requirements of the inspection due to high cost and low precision.Through the in-depth study of machine vision inspection technology,building a detection platform based on machine vision technology.From hardware selection,vision system design,algorithm research and camera calibration and other aspects of research.Based on this,the parallelism of the mechanical parts,Verticality,Concentricity,and Shape defect detection of the part as the research objectives.And then,selecting suitable test objects and conducting research on them.Finally,All tests meet the use requirements.Research on parallelism and perpendicularity error detection of mechanical parts.Selecting the block as the experimental object.First,the edge to be detected is extracted,and the real edge is obtained by fitting the detection edge with least square method.Then the angle between the edges to be measured is calculated,and the angle deviation is obtained,and then the parallelism and perpendicularity error of the target to be detected are multiplied by the corresponding length.From the measured data,the measurement accuracy is more accurate than the traditional method.Research on the coaxiality detection for shaft parts.Firstly,studying the algorithm of the shaft parts detection,and establishing the model of shaft part coaxiality detection based on machine vision.Then acquiring the image of the step shaft and processing the image and extracting the edge contour of the image.Then divide the cross-section of the stepped shaft,extracting the data points of the cross-section projected on the edge of the image,according to the principle that the intersection point of the vertical bisector of the two circle segments of the cross-section circle is the center point of the coordinate,the staircase axis can be obtained,the coordinates of the center of each circle on the cross section.The least squares fit of the coordinates of the circle center point of the reference axis to obtain the equation of the reference axis;Finally,the distance between the circular center coordinates of each section of the measured axis and the equation of the reference axis is calculated,and comparing the size of each distance,and obtaining the coaxiality error of the staircase axis.Research on defect detection for parts,selecting involute gear as experimental object.By comparing the detected gear image with the standard gear image and using the relevant image processing and recognition algorithm,realizing the tooth profile defect detection.Firstly,the image of the standard involute spur gear is acquired by Halcon software and the dimension detection is carried out to obtain the basic parameters of the gear;then the standard involute tooth profile is drawn according to the obtained gear parameters;Finally,the Hausdorff distance between the actual tooth profile and the standard tooth profile is calculated,and based on this,we can determine whether the gear is qualified.This paper extracts the actual gear profile and seeks the Hausdorff distance from the standard gear tooth profile to detect gear tooth defects.The example proves that this method can effectively detect tooth defects and providing an effective detection method for the detection of gear tooth defects.