Grinding Wheel Wear State Recognition Based On Light Scattering Simulation And Machine Vision

With the rapid development of precision and ultra-precision industry in China,the demand for precision machining tools represented by diamond grinding wheel is gradually increasing.In order to accelerate the construction of a new digital transformation pattern of precision and ultra-precision machining in China,and drive and promote the realization of intelligent monitoring system of precision grinding,there is an urgent need for a fast and convenient diamond grinding wheel wear state recognition method.Therefore,on the basis of summarizing the limitations of indirect methods such as acoustic emission and grinding force,this paper studies the direct methods represented by machine vision.By analyzing the problems of unclear physical significance and time-consuming and laborious sample acquisition in the current machine vision methods,a grinding wheel wear state recognition method based on light scattering simulation and machine vision is proposed.This method has certain theoretical significance for studying the relationship between grinding wheel wear and its image,and has certain application value in realizing non-contact and high-precision recognition of grinding wheel wear.With the support of the Key project of intergovernmental International Science and Technology Innovation Cooperation(2017YFE0128400),the recognition method of grinding wheel wear state based on light scattering simulation and machine vision is deeply studied,a simulation method of grinding wheel surface light scattering phenomenon is proposed,the image feature index is designed and verified,and a set of grinding wheel wear state recognition system based on support vector machine and gray level co-occurrence matrix is preliminarily established.The main research work of this paper is as follows:(1)The characterization and simulation methods of engineering surface topography are summarized.According to the measured statistical parameters of grinding wheel surface,the simulation morphology of worn grinding wheel is obtained by using two-dimensional digital filtering technology;The surface of the simulated grinding wheel is processed by low-pass filtering,and the number of abrasive particles on the surface is calculated.(2)The physical process of light scattering imaging from the surface of grinding wheel is introduced.Based on the method of wave optics,the near-field model of light scattering on the surface of grinding wheel is established,and the energy difference of light scattering on the surface of grinding wheel’s grain and bond is analyzed;Based on the method of geometric optics,the light scattering far-field models of six kinds of worn wheel surfaces are established,and the relationship between the scattered energy of wheel surface and the surface height and the number of abrasive particles is analyzed.(3)The experimental scheme of grinding wheel surface imaging is designed,and the surface images of six kinds of worn grinding wheels are obtained;an image feature index based on binary image brightness is proposed.By comparing the simulation energy index of light scattering on the surface of grinding wheel,it is found that the change law of the two is similar,which verifies the effectiveness of the light scattering simulation method.(4)Based on support vector machine and gray level co-occurrence matrix,a grinding wheel wear state recognition method is proposed.Six kinds of worn grinding wheels are recognized,and the recognition effect is evaluated by confusion matrix.The results show that this method has high recognition accuracy,which is reached 96.67%,for the wear wheel image.