Cutting Tool Wear Region Extraction Based on L0 Gradient Minimization Model and Improved CV Model

To accomplish the wear region extraction in the cutting tool wear detection system exactly, an image segmentation method based on L0 gradient minimization model and improved CV model is proposed in this paper. Firstly, the original image is smoothed by L0 gradient minimization model. Then the Otsu method is used to segment the smoothed image to achieve the initial location of the wear region in the cutting tool surface, after which the one-pixel edges are obtained by using the edge following algorithm. Finally, by taking the edges as initial contour, the CV model is applied to segment the original image to extract the wear region exactly. Meanwhile, to reduce the long runtime caused by the low convergence speed, an acceleration factor based on the gradient is introduced to optimize the CV model. A lot of experimental results show the proposed method has a good performance in cutting tool wear region extraction, with high speed and high precision.