Machine learning-enhanced vision systems for cutting tool notch detection in new energy battery manufacturing

This paper presents a study on the problem of burrs on the electrodes of new energy batteries, which are a major factor contributing to battery short-circuits and explosions. During the process of electrode cutting, the use of cutting tools with a notch is likely to cause burrs on the electrode. Therefore, it is essential to accurately detect the notch of the cutting tool. This paper explores the issue of cutting tool notch detection using machine learning-enhanced vision systems. Firstly, a set of cutting tool image acquisition devices is used to capture high-quality images of the cutting tool edge. Next, an algorithm for removing attachments based on concave point matching is proposed, effectively eliminating edge attachments by analyzing the concave point information of the edge. Additionally, we propose an enhanced Zernike moment sub-pixel edge extraction method, which achieves sub-pixel edge extraction while preserving the edge characteristics of the cutting tool. Furthermore, a notch detection algorithm based on quartic Hermite interpolation is introduced to detect the notch of the cutting tool by initially identifying the tool’s edge. The proposed algorithms are compared with other state-of-the-art methods, demonstrating faster and more accurate extraction of sub-pixel cutting tool edges and detection of cutting tool notches.