Robust Features Extraction for Structured Light Images of Welding Seam in Multi-pass Submerged Arc Welding

摘要

A robust method of image processing and features extraction is presented for structured light images of welding seams in multi-pass submerged arc welding. Firstly, the ranges of welding seams in captured images are detected by distribution of intensity search. Secondly, the profiles of grooves are obtained with the method of column based intensity distribution. Then the main characteristic vectors of profiles are obtained using Hough transform. According to different features of profiles in multi-pass welding, a center of gravity detection method is presented to extract feature points on the basis of conventional corner detection method. Finally, experimental results testified the effectiveness and robustness of the algorithm.