Research On The Technologies Of Gear Multi-parameter Measuring Based On Machine Vision

As an indispensable part of the mechanical transmission,gear has been always one of the main science research objects for researchers and engineering personnel.In this paper,the machine vision and image processing technology were applied to obtain the values of gear parameters and the measurement error detection projects.The camera calibration technique,pixel level edge detection methods and sub-pixel edge detection technology based on the image gray gradient,positioning technology of the gear center and the measurement technology of important parameters of gear and error detection deviation measurement technology were analyzed in this paper at length.Based on the measurement platform,the MATLAB software was used to realize the measurement and acquisition of the deviation values of the parameters and error detection items of the involute standard spur gear.Firstly,the hardware structure and software function structure of gear parameters measurement platform with the backlight were completed.The calibration of the camera was used to calibrate the camera’s internal and external parameters and correct image distortion.Secondly,by contrasting pixel level gear edges detected by a variety of edge detection algorithms which were commonly used,the edge figure which had the least false edge and the best stability detected by the LoG operator was selected out for the subsequent operations.On the basis of the gear teeth edge image isolated from the gear edge graph by using the mathematical morphology,a completed and closed profile of the figure was obtained after removing noise point groups and connecting the tooth profile on the breakpoints in the graph.Thirdly,according to the distribution law of gray value of pixels in the image,an algorithm was established to calculate the gray value of each equivalent pixel in the normal direction of the closed edge and the equivalent distance of each equivalent pixel to the center point.On the basis of analyzing existing sub-pixel edge detection algorithms,a method obtaining the sub-pixel edge position by interpolating the gray gradient values of the equivalent pixel in the normal direction of the edge and fitting the interpolation results in the longitudinal direction of the edge was proposed.Then,with the help of the gray barycenter method,the center of the gear was coarsely located,and according to the change of the distance values of each point on the tooth profile to the center of the gear which were coarsely located formerly,the pixel points on the top of the tooth circle were selected to be the basement of the precise positioning of the gear center.By comparing the distances between the center of the gear and the pixels on the gear tooth profile calculated by three-point method,the least square fitting method and the method proposed in this paper,it was proved that the location of the pixel points on edge got by sub-pixel edge detection method proposed in this paper is more accurate.Finally,based on the gear design process,the measurement order of gear parameters was determined.According to the result of image calibration,the basic parameter values of gear such as the number of teeth and the modulus were obtained.By utilizing the edge detection method proposed by this paper and a previous interpolation fitting method proposed by other one,the sub-pixel coordinate values of pixel points on the left and right tooth profiles segmented before were obtained.Then the pitch deviation of gear,the tooth profile deviation and common normal deviation were measured.According to the precision grade of the tested gear and the comparison of the results obtained by the two methods,the sub-pixel edge detection method and the gear parameters measurement method proposed in this paper were proved to be feasible and effective.