Study On The Measurement Of Surface Roughness And Surface Defects Using Optical Method

The surface quality of steel has a great influence on the performance and life of its products.Surface roughness and surface defects are important indicators for measuring surface quality.With the increase of social productivity,the demand for high-quality steel is increasing,and the detection of steel surface roughness and surface defects is becoming more and more important.Until now,most of the surface quality measurement methods are limited to measure surface roughness or detect surface defects,and mostly focus on off-line measurement.Few paper describe the technology for measuring simultaneously surface roughness and surface defects.In this paper,a system for surface roughness and surface defect measurement is built.The laser light scattering technology is used in this system to measure the surface roughness,and the measurement ranges from 0.025?m to 0.8?m.And machine vision technology is used to detect surface grooves and bumps of the workpiece.By analyzing the acquired surface scattering image,three characteristic parameters of discrete variance S,bright point ratio BPR and bright gray ratio BGR are proposed to evaluate the surface roughness of the workpiece,and the relationship between these parameters and surface roughness can be given by fitting.Meanwhile,BP neural network,support vector regression machine(SVR)and convolutional neural network(CNN)are used to predict surface roughness.The predicted root mean square errors of BP neural network and SVR are 0.0012 and 0.0192,respectively.And the accuracy of classification using CNN model is 100%.Image processing algorithms are proposed to find and locate the defects on the workpiece surface,and also used to calculate their sizes in the measurement direction.Meanwhile,the feasibility of laser as a light source to detect surface defects is verified.The surface grooves with a width 0.2mm to 1mm and the surface bump with a diameter of 3mm and 4mm are measured.And the measured root mean square errors are 0.0122 and 0.2458,respectively.A transparent window device is designed to realize on-line detection of surface parameters using Solid Works.And Flow Simulation module is used to simulate the fluid flow.The designed transparent window device is applied to measure the surface roughness,and the simulation results and the feasibility of the device can be verified.Experimental results show that this device can be used for on-line measurement of workpiece surface roughness under coolant condition.This study provides a basis for on-line detection of surface roughness and surface defects of workpieces.