Improved GrabCut Algorithm Based On RANSAC For Tire Belt Image Segmentation

Tire quality is the foundation of tire production. With the increase of tire output, defect detection of tire forming process becomes more and more important. In the process of tire forming, the segmentation of the tire belt is the basis of detection. In this paper, a method combining RANSAC algorithm and GrabCut function is proposed to segment the tire belt. Firstly, the surface fitting is carried out by RANSAC algorithm, the drum surface is corrected to a plane and the difference between the drum plate and the drum seam is reduced at the same time. Then, the fixed rectangular frame is set up in combination with GrabCut algorithm to segment the belt, and the small connected domain is removed after the image binarized to process the results. Finally, the results are compared with the traditional GrabCut algorithm and U-net algorithm. The comparison results show that the algorithm in this paper can segment the tire belt image more accurately and effectively without the need for human interaction.