Research On Gear Surface Roughness Measurement Methods Based On Machine Vision

The Gear surface roughness is closely related to the wear resistance,fatigue strength,fit properties,corrosivity and energy consumption of gear pairs.The surface roughness affects the performance and life of the machine’s transmission directly.However,owing to the low efficiency and the limitations of the overall evaluation of the surface,the current methods cannot meet the requirements of automated on-line measurement.At present,the measurement methods based on machine vision have the advantages of high efficiency,large measuring and non-contact.And the methods mostly study planar objects,with less research on curved surfaces.Therefore,this research has explored and studied the measurement of the gear surface roughness by means of machine vision,and has a certain engineering value to achieve accurate,efficient,on-line measurement.In view of the current status of gear surface roughness measurement,this paper proposes a measurement method of gear surface roughness based on machine vision and studies the imaging mechanism and feature extraction methods.A semi-automated gear surface roughness measurement system was preliminarily established and the effectiveness of the method was verified.The main research works of this paper are as follows:(1)The definition and evaluation of gear-tooth surface roughness are introduced,and the characteristics of surface roughness are studied.The mechanism of gear-tooth surface imaging was analysed from the light scattering principle,and the correlation between surface images and roughness of different roughness levels was explained.And,the digital image processing technology required for edge extraction is introduced in detail.Finally,this research work preliminarily establishes the detection scheme of gear-tooth surface roughness based on machine vision,and introduces the image acquisition module and image processing module in the measurement system in detail.(2)A method for measuring tooth surface roughness of gears based on shape features is proposed.From the principle of light scattering,the shape profile in the tooth surface image is extracted.And an evaluation index algorithm for the feature index is designed from the degree of scattering,structural information,and intensity information.Then,a experimental device is designed according to the detection scheme.The experimental results show that there are obvious correlations between gear-tooth surface roughness and structural information(Eigmax,Eccentricity and Area)and strength information(E).The accuracy,monotonicity,and stability of the proposed metrics were evaluated.(3)Using the method of support vector machine(SVM),the prediction model between gear-tooth surface roughness and image feature index was established to achieve high-precision prediction of gear-tooth surface roughness.The cross-validation method is used to optimize the parameters in SVM.Finally,the predictive model is tested and verified.The evaluation results show that this method improves the prediction accuracy of gear-tooth surface roughness.(4)Based on LabVIE W and MATLAB,a gear surface roughness measuring system based on machine vision was developed.Capture image In LabVIEW and call MATLAB program to achieve image processing.Then design the main function module of gear-tooth surface roughness measurement system.Finally,the response time and the run time of the measurement system were tested.The results show that the measurement system has the advantages of high efficiency and stability and can measure the gear-tooth surface roughness in real time.