A machine vision-based radial circular runout measurement method

In order to overcome the disadvantages of contact measurement, time-consuming in traditional radial circular runout measurement, a non-contact measurement method based on machine vision technology is proposed in this paper. Using the image template matching and sub-pixel edge extraction techniques, the fine edge information of measured region is obtained. The rand sample consistency (RANSAC) algorithm is performed to filter the noise of edge data. Then, the least squares method is used for the parallel line fitting of two edges of measured region, and the reference axis of the shaft is gotten. The distance of measured point to the reference axis is defined as the radius at this rotating position, and the radius fluctuation at different rotational positions is the radial circular runout. The proposed method also allows random rotation of the measured shaft during the measuring process, and the high-precision-assisted rotation equipment is not essential, which enhances the convenience of the measuring equipment.

Automated summary

This paper proposes a non-contact measurement method based on machine vision technology to overcome the disadvantages and time-consuming issues in traditional radial circular runout measurement. The fine edge information of the measured region is obtained using image template matching and sub-pixel edge extraction techniques. The RANSAC algorithm is applied to filter the noise in the edge data. The least squares method is used to fit the parallel lines of the measured region, obtaining the reference axis of the shaft. The distance from the measured point to the reference axis defines the radius at the rotating position, and the fluctuation of radius at different rotational positions is the radial circular runout. This method allows random rotation of the measured shaft during the measuring process and does not require high-precision-assisted rotation equipment, enhancing the convenience of the measurement equipment.