On-machine Measurement Of Workpiece Dimensions Based On Binocular Vision

Under the requirements of the "intelligent manufacturing" era,the use of machine vision-based measurement technology to achieve on-machine measurement of metal workpieces had important practical application values for improving detection efficiency,processing efficiency and accuracy.In the current machining process,the workpiece needed to be disassembled during measurement,and then clamped and repositioned twice,which caused problems such as low efficiency and accuracy when manually measuring dimensions.Therefore,the current manual measurement of the size cannot complete the measurement task with high efficiency.Today,with the continuous development of high-tech industries,enterprises need a method for real-time on-machine measurement of workpiece dimensions,and they need to accurately measure key dimensions.Machine vision technology was developing rapidly,and visual measurement technology was being widely used in different production fields.Vision measurement technology had the advantages of non-contact and no damage to the surface of the workpiece,faster measurement,better accuracy.And it provided technical support for non-contact on-machine measurement.This thesis was based on binocular vision measurement technology,according to the characteristics of the on-machine measurement environment and the size of the metal workpiece,researched on image acquisition,binocular vision system calibration,workpiece image highlight processing,stereo matching and other important methods.The on-machine measurement platform based on the machine tool was built to realize the on-machine measurement of the key dimensions of metal workpieces.First,based on the characteristics of the binocular vision measurement system and the measured workpiece size,the hardware of the measurement platform was selected,including the camera lens,camera bracket,camera fixing plate and data cable,and the off-machine measurement platform was set up.Then,the binocular vision system calibration based on Zhang Zhengyou’s plane calibration method was introduced,and calibration experiments were performed,the internal and external parameters of the system were obtained,and the calibration results were accurate.In terms of image processing,aiming at the problem that high-reflectivity workpieces produced high-light areas due to uneven illumination distribution and other factors,this problem affected the feature extraction and stereo matching of the workpiece images,so a method for high-light processing of metal workpieces based on the PSO-OTSU algorithm was proposed.Experiments showed that the influence of high light area on measurement accuracy can be reduced.In terms of stereo matching algorithm,an improved feature point matching algorithm based on epipolar correction was proposed,which mainly included epipolar correction,feature point extraction,matching search and mismatch optimization.Comparison of experimental results showed that this algorithm can obtain more feature point matching pairs of coordinates,and provided data for subsequent 3D coordinate calculation of feature points.Based on the above research,according to the processing technology of the measured workpiece and the characteristics of the machine tool,an on-machine measurement platform based on binocular stereo vision was built.At the same time,the method of calculating the three-dimensional coordinates of the feature points and the size of the workpiece was studied and the measured results were compared with the measured results of the three coordinate measuring machine to determine the factors that affected the measurement accuracy,and the experimental verification was carried out,which provided a basis for subsequent control errors.By analyzing the measurement results of repeated experiments and comparative experiments,the visual measurement system can better calculate the size of the metal workpiece,and realized the on-machine measurement of the workpiece size.