Research on Roundness Measurement Method for Thin-Walled Metal Tube Ends Based on Machine Vision

In modern manufacturing, the roundness of thin-walled metal tube ends is a critical parameter in industrial quality control. However, traditional contact-based methods for measuring metal tube end roundness have limitations, such as long detection times and cumbersome processes, making them unsuitable for automated production. In response to this, this paper focuses on the thin-walled metal tube ends, constructing an image acquisition system for the tube end surfaces based on machine vision and conducting image preprocessing. First, the Canny operator is employed for initial edge detection, followed by sub-pixel edge localization and extraction using Zernike moments along the gradient direction. Finally, the least squares method is applied to fit a circle to the contour of the metal tube end, and the roundness is calculated based on the fitted circle. The experimental results indicate that the roundness measurement using machine vision achieves a relative error within ± 5% and a detection accuracy of up to 10 µm. These findings confirm that the proposed method satisfies the stringent requirements for non-contact roundness measurement of thin-walled metal tube ends, demonstrating its precision and suitability for industrial applications.