Research On Rapid Detection Of Stepped Shaft Base On Binocular Vision

As one of the typical parts often encountered in mechanical devices,the accuracy of shaft parts is directly related to the quality and service life of mechanical transmission parts.Most of the current measurement methods still rely on the traditional way of manual detection.This measurement method is easy to cause part surface scratch when using the detection device,and it is difficult for operators to maintain high accuracy and low error rate in the long-term measurement process.With the proposal of industrial intelligence and automation,the traditional machinery industry continues to integrate with other technologies,resulting in a non-contact measurement method based on machine vision,which has begun to gradually replace manual detection.This paper focuses on the high-precision non-contact measurement of stepped shaft,and proposes a rapid detection method of stepped shaft based on binocular vision.The main research contents are divided into the following parts:(1)Aiming at the problem that the traditional edge extraction algorithm is easy to be affected by light,so it is difficult to obtain the contour of metal parts,a contour detection algorithm combining image segmentation and gray value gradient change is proposed,which transforms the edge detection problem into image segmentation problem.By analyzing the gray histogram of the image,the optimal segmentation threshold is found,and then the image is binarized.Finally,the part contour is extracted through the gradient change of gray value.(2)Based on the parallax principle of binocular vision,an axis diameter measurement model based on binocular vision is constructed.The model calculates the center coordinates of both ends by fitting the straight line of the thimble at both ends,takes the center as the rotation point,uses the non parallel binocular vision model to calculate the three-dimensional coordinates of the contour point,and then uses the contour point and rotation point to calculate the center point coordinates of the contour surface,so as to calculate the axis radius.(3)In order to realize the geometric tolerance measurement of shaft parts,the coaxiality error measurement model and cylindricity error measurement model of shaft parts are established respectively.The shaft parts are rotated by stepping motor,and multiple groups of part images taken by camera are used as measurement data.Aiming at the problem that the coordinate system of contour points under different rotation angles does not belong to the same coordinate system,a spatial coordinate rotation model is proposed,and then all contour points on the same contour surface calculated by the model are brought into the minimum distance objective function to calculate the center of the contour surface.Finally,the coaxiality error and cylindricity error are calculated respectively according to the minimum double evaluation method.(4)According to the above technology,take the stepped shaft as the experimental object to measure and verify the feasibility of the method,then build the step axis measurement system based on binocular vision,and carry out the system visual measurement experiment.The experimental results show that the maximum measurement error of stepped shaft radius is controlled within 0.14 mm,and the maximum errors of coaxiality and cylindricity are controlled within 0.13 mm and 0.35 mm respectively,which proves that the proposed non-contact measurement method is feasible.