Search

Measurement technology based on machine vision has been widely used in various industries. The development of vision measurement technology mainly depends on the process of photosensitive components and the algorithm of processing a target image. In the high-precision dimension measurement of machined metal parts, the high-resolution imaging device usually exposes the cutting texture of the metal surface and affects the accuracy of measurement algorithm. At the same time, the edges of machined metal parts are often chamfered, which makes the edges of objects in the picture overexposed in the lighting measurement environment. These factors reduce the accuracy of dimensioning metal parts using visual measurements. The traditional vision measurement method based on color/gray image makes it difficult to analyze the physical quantities in the light field except for the light intensity, which limits the measurement accuracy. Polarization information can more carefully depict the edge contour edge information in the scene and increase the contrast between the foreground and the background. This paper presents a method to improve the measurement accuracy of machined metal parts by using polarization vision. The incident angle of the light source is optimized according to the complex refractive index of the metal material, and the degree of polarization image with enhanced edge contour features of the ROI (region of interest) is obtained. The high-precision measurement of cylindrical brass motor components is realized by using the method of reprojection transformation correction and maximum correlation template matching (NCC) for rough positioning, as well as the method of edge extraction and optimal fitting. The experimental results show that for copper parts with a tolerance range of ± 0.1 mm, the average measurement error and maximum measurement error are within 0.01 mm, which are higher than the existing color/gray image measurement methods. [ABSTRACT FROM AUTHOR]

Measurement technology based on machine vision has been widely used in various industries. The development of vision measurement technology mainly depends on the process of photosensitive components and the algorithm of processing a target image. In the high-precision dimension measurement of machined metal parts, the high-resolution imaging device usually exposes the cutting texture of the metal surface and affects the accuracy of measurement algorithm. At the same time, the edges of machined metal parts are often chamfered, which makes the edges of objects in the picture overexposed in the lighting measurement environment. These factors reduce the accuracy of dimensioning metal parts using visual measurements. The traditional vision measurement method based on color/gray image makes it difficult to analyze the physical quantities in the light field except for the light intensity, which limits the measurement accuracy. Polarization information can more carefully depict the edge contour edge information in the scene and increase the contrast between the foreground and the background. This paper presents a method to improve the measurement accuracy of machined metal parts by using polarization vision. The incident angle of the light source is optimized according to the complex refractive index of the metal material, and the degree of polarization image with enhanced edge contour features of the ROI (region of interest) is obtained. The high-precision measurement of cylindrical brass motor components is realized by using the method of reprojection transformation correction and maximum correlation template matching (NCC) for rough positioning, as well as the method of edge extraction and optimal fitting. The experimental results show that for copper parts with a tolerance range of ±0.1 mm, the average measurement error and maximum measurement error are within 0.01 mm, which are higher than the existing color/gray image measurement methods.