Research On Visual Measurement Of Roughness Based On Electromagnetic Scattering Simulation

The precision and intelligence of the manufacturing industry has increasingly become a major trend and core content for the future development of the manufacturing industry.The processing method represented by grinding is an important means to achieve precision manufacturing.Surface roughness is closely related to many properties of mechanical parts such as lubrication,coordination,vibration and service life.Machine vision-based roughness measurement is expected to achieve online or in-situ measurement of workpiece surface roughness during precision grinding due to its many advantages such as high integration,strong scalability,and good stability,thereby improving processing quality and intelligence.The current research on the visual inspection of roughness has the problems of time-consuming and laborious to establish the experimental prediction model,the unclear mechanism of the scattering imaging of the workpiece surface,and the lack of the design optimization method of the image feature index.In view of the importance of visual measurement of surface roughness to improve the quality and intelligent level of precision manufacturing and the lack of existing research work.Taking grinding as an example,this paper proposes a visual measurement of roughness based on electromagnetic scattering finite element simulation,and compares and verifies with experiments.The main research work is as follows:(1)Aiming at the problem of time-consuming and laborious establishment of roughness visual measurement experimental prediction model,a method of roughness visual measurement research based on electromagnetic scattering finite element simulation is proposed.The principle and implementation process of various commonly used computational electromagnetic methods are studied,and the derivation process,advantages and disadvantages,applicable conditions and limitations of Kirchhoff approximation method,perturbation method,moment method and finite element method are analyzed.(2)Aiming at the problem of unclear meaning of image characteristic indexes due to the unknown surface scattering imaging mechanism of the workpiece,an object surface scattering model is established to explain the influence of object surface information on the distribution of the scattering field.Based on the principle of photometry,the light scattering imaging process is analyzed,and the energy conversion in the process of light from the object plane to the image plane is quantitatively analyzed.The photoelectric conversion model of CCD sensor is established,which explains the conversion process of information from optical signal to electrical signal to digital signal.Through the above analysis,the relationship between the pixels of the digital image and the surface roughness is interpreted,and the analysis of the actual imaging process is simplified.The one-dimensional rough surface of grinding surface and the corresponding finite element simulation model are established,and the solution analysis of the model is completed.(3)Aiming at the problem of the lack of image feature index design optimization methods,based on the results of finite element simulation,it is proposed that the characteristics of the scattered field such as the energy size,energy information distribution,and approximate coincident area of energy are related to the roughness.Designing the energy index E on the monochromatic background,the energy sum index EP and the energy difference index EL on the two-color background,the energy information entropy index H on the monochromatic background,and the coincident area index S under the incident of blue and red bicolor light.The influence of the approximate definition of regional energy on the coincidence area index S is analyzed,and the design and optimization of the coincidence area index S are realized.(4)In order to verify the effectiveness of the research method of roughness visual measurement based on electromagnetic scattering finite element simulation,a corresponding roughness visual measurement experimental system is built.Experimental images are used to verify the correlation between the image feature indexes E,EP,EL,H,S and roughness under various color backgrounds,and the S index goodness of fit R~2 was improved from 0.8269 to 0.9478.Compared with the Ga index and F2 index designed by the original method,the index designed by this research has outstanding advantages,and its fitting accuracy is significantly higher than the index designed by the original method.The research work of the thesis provides an important reference for the visual measurement of surface roughness.To a certain extent,it helps to promote the application of machine vision in the field of online measurement of roughness,thereby improving the processing quality and intelligent level of precision manufacturing.