An In-Situ Tool Wear Measurement Method Based on Super-pixels and Enhanced Corner Detection Algorithm

Abstract In robotic milling scenarios, lighting conditions that are difficult to optimize and random angles of the cutting edge often lead to inaccuracies in in-situ tool wear detection. This paper proposes a method based on machine vision to accurately measure the wear parameters of end mills. Firstly, a cropping algorithm based on polar coordinate transformation is introduced to extract the Region of Interest (ROI) from the flank face of the tool bottom. Subsequently, a segmentation algorithm that combines super-pixels and multi-thresholding is proposed, with a focus on maintaining the connectivity of the damage regions during segmentation. Finally, an algorithm based on improved corner detection is presented to reconstruct the tool breakage region. Experimental results demonstrate that the proposed method achieves precise tool wear measurement despite random cutting edge angles and exhibits good robustness under varying lighting conditions. In the in-situ detection experiment, the method achieved a measurement accuracy of 97% for maximum wear width and 95% for wear area, indicating the effectiveness of the proposed method.