A new seam location extraction method for pipe-line backing welding of MAG based on passive optical vision sensor

Seam tracking is a prerequisite of implementing automatic welding process. Gap variance has an important impact on the welding quality of MAG for pipe-line backing welding. A welding process monitoring system based on CCD is designed. However, the images obtained from CCD with MAG welding often have much noise. Therefore, accurate and high efficient image processing methods for extracting seam location are urgently needed. A new seam location extracting method is presented, the molten welding pool area is obtained after image preprocessing, and then the Sobel transformation method is adopted to choose the initial contour control point for the greedy Snake model. Finally, the greedy Snake model was used to fit the welding seam location. The experimental results show that, the proposed method combined with Sobel transformation and greedy Snake model is also effective to the images with much noise.