Research On Gear Geometry Parameter Detection Method Based On Machine Vision

Gears are the most widely used components in industry,with the characteristics of smooth transmission and long-lasting durability.Their manufacturing level and accuracy have a significant impact on the performance of mechanical products.Modern industry has put forward higher requirements for the accuracy of gear production and manufacturing,and improving gear measurement technology and improving the measurement accuracy of gear geometric parameters is extremely important for improving the level of gear production and manufacturing.In response to the problems of high difficulty in gear measurement,complex measurement devices,and low automation level,this article takes straight cylindrical gears as the research object,conducts in-depth research on image processing technology based on machine vision and gear geometric parameter measurement and evaluation methods,and achieves non-contact measurement of straight cylindrical gears.The research content of this article is as follows:Firstly,based on the measurement process of gear parameters,the overall design of a gear geometry parameter measurement system consisting of an image acquisition module and a software module was completed.Analyze the hardware parameters and requirements of the system,determine the selection of hardware such as camera,lens,and light source,and select the lighting mode of the light source to achieve the shooting function of the system.Using a checkerboard calibration board as the object for camera calibration,complete lens distortion correction and pixel equivalent calibration.Secondly,by studying the changes in image edges under different lighting intensities,a gear edge error compensation algorithm based on lighting intensity is proposed.Based on the detailed features of gear images,key algorithms for image processing were studied,and the preprocessing of gear images was completed.By studying traditional edge detection algorithms and improving the Zernike moment subpixel edge detection algorithm,the Canny operator is used to coarse locate edge points,and the maximum inter class variance method is used to automatically select the optimal threshold.The accuracy of the subpixel detection algorithm in this paper is verified through experiments.Finally,study the measurement methods of gear geometric parameters,design a set of gear geometric parameter measurement algorithms,and obtain the actual parameters of the gear based on the calibration results.Firstly,use the three-point rounding method to accurately locate the gear center.Then,based on the principle of gear design,complete the measurement of basic geometric parameters such as tooth number,tooth tip circle radius,tooth root circle radius,gear module,indexing circle radius,and base circle radius in sequence.Finally,based on the measurement results of basic geometric parameters,the measurement of gear pitch deviation and the change in common normal length is completed.Analyze the measurement accuracy of the gear detection system by comparing the measurement results with the actual size of the spur gear.The measurement results of the detection system for the diameter of the gear tooth tip circle and the diameter of the tooth root circle have errors below 0.03 mm,with an error rate within 0.02%.The measurement results of tooth pitch deviation and common normal length variation meet the accuracy level requirements of level six,respectively.The maximum values of each deviation are within a reasonable range.This proves the rationality of the algorithm for detecting gear geometric parameters in this paper.