Workpiece recognition and positioning for laser marking machine based on machine vision

Abstract

In intelligent laser marking, workpiece positioning accuracy directly affects marking quality. Traditional mechanical fixture positioning struggles to adapt to complex working conditions such frequent workpiece changes and random positional offsets, which often leads to marking deviations. To address this issue, this paper proposes a machine vision-based method for workpiece recognition and localization. First, an integrated contour extraction pipeline comprising enhanced bilateral filtering, adaptive Otsu thresholding, morphological optimization, and improved Canny edge detection is constructed to enhance edge extraction accuracy under complex conditions. Subsequently, precise tooth tip recognition and subpixel localization of the inner hole center are achieved, providing high-precision pose parameters for laser marking. On a test set with illumination disturbances and salt-and-pepper noise, the proposed algorithm achieves a recognition accuracy of 98.53%, a localization error of, and a processing time of per image, significantly outperforming comparison algorithms while meeting real-time requirements. Generalization experiments verify its robustness and versatility in complex industrial scenarios, providing key technical support for the intelligent upgrading of laser marking equipment.