Monocular line scan vision-based surface defect detection approach for highly reflective bearing balls

Surface defects strongly affect the stability and service life of bearing balls. In this Letter, I present a monocular line scan vision-based detection system for detecting surface defects on bearing balls. An optical system was designed to solve the problems of nondevelopability, large spherical curvature, and high reflection of bearing ball surfaces. The principle of light illuminating bearing balls was developed. By analyzing the motion unfolding trajectory curve, I propose a line scanning unfolding process and image acquisition scheme for the whole surface of the bearing ball. According to the unfolding principle, I have established a mathematical model of the whole-surface bearing ball unfolding process and developed a simulation. Experiments were performed to capture the surface image of bearing balls. A defect detection algorithm for spatiotemporal image is developed. A subtraction operation is used to enhance the defect information. Spatial-temporal resolution normalization is developed to make the scale of spatiotemporal image uniform and extract the surface defects. The experimental results show that the detection resolution of the crack defects is approximately 0.001 mm2, and the crack defect detection rate is 100%, which demonstrates that the proposed method has high detection accuracy. (c) 2025 Optica Publishing Group. All rights, including for text and data mining (TDM), Artificial Intelligence (AI) training, and similar technologies, are reserved.