Image Segmentation of the Underwater Welding V-seam Image Based on Improved Two-dimensional Ostu Algorithm

摘要

To meet the request of V-seam tracking of underwater welding robot vision sensor, an improved two-dimensional image segmentation algorithm Ostu is presented in this paper. According to this algorithm, changing discriminate domain partition of 2D histogram, introducing a relaxant variable and enlarging the searching step of threshold we get the optimal threshold. The experimental result shows that the computing time of this algorithm is shorter than the traditional and anti-interference capability is much stronger. It can meet the requirements of the underwater welding seam tracking.