Research On Measurement Technology Of Sprocket Radial Runout Based On Machine Vision

The sprocket is the main part of the chain drive,and its processing accuracy directly affects the working performance of the chain drive.Radial runout is an important part of the shape and position error,and is one of the main indicators for evaluating the machining accuracy of the sprocket,which directly affects the dynamic characteristics of the chain drive.Measuring the sprocket radial runout is an important prerequisite for improving the processing accuracy and assembly accuracy of the sprocket and ensuring the quality of the sprocket transmission.With the advancement of modern automobile engine manufacturing technology,online or in-situ measurement of sprocket radial runout is required.In this paper,the in-situ or online measurement of sprocket radial runout is used as the goal.Using machine vision measurement technology,a machine vision measurement model of sprocket radial runout is established.Since the geometry and measurement parameters of the sprocket are similar to those of shaft and tooth parts,the method of establishing this model can also be used to study the visual measurement of related parameters of other shaft and tooth parts.Firstly,based on the two-step calibration method based on the plane,an improved camera calibration algorithm for correcting the main point of the camera is proposed.Using the shortest line segment set between two pairs of conjugate curves,the coordinates of the vertex of the polygon formed by the intersection of these line segment sets are obtained,and the camera principal point is corrected by solving the polygon centroid.The size of the standard gauge block is measured,and the accuracy of the camera calibration is improved through experiments to improve the accuracy of the algorithm,which lays the foundation for improving the accuracy of visual measurement.Secondly,considering the influence of factors such as the lighting conditions of the machinery factory,the color of the sprocket surface and the scene background,based on the analysis of the existing sub-pixel edge detection algorithm,the accuracy and noise resistance of the detected edge are used as indicators.The experiment tested the edge interpolation algorithm,the spatial moment method and the fitting method,and determined the subpixel edge detection algorithm suitable for extracting the sprocket edge image on the spot.Thirdly,a world coordinate system is established on the end surface of the sprocket to describe the sprocket under test.Place a special concentric ring in the center hole of the sprocket,and take an image of the end face of the sprocket.Determine the transformation relationship between the world coordinate system and the camera coordinate system using the known ring radius and sprocket center hole radius and the coefficients of the curve equation of the ring and the sprocket hole edge obtained by ellipse fitting,that is,external parameters for visual measurement of sprocket radial runout.That lays the foundation for the establishment of a visual measurement sprocket radial runout model.Then,using external parameters to visually measure the radial runout of the sprocket,the detected sprocket edge point is projected onto a coordinate plane parallel to the surface of the sprocket,and ellipse fitting is performed on the tooth groove edge point and the tooth tip edge point,respectively,to solve the lowest point of the groove arc and the highest point of the tooth tip arc.At last the radial runout measurement model of the tooth groove and the tooth tip relative to the sprocket center point is established.Finally,on the basis of the machine vision measurement model established in this paper,the radial runout of four sprockets with different pitches and number of teeth is measured and compared with the measurement results of the mapping projector and the coordinate measuring machine to verify the correctness of the machine vision measurement model and then analyze the factors that affect the measurement accuracy.