The Study On Image Recognition Arithmetic Of SAW Seam Self-tracking System

At present, welding manufacture is developing its full-automatic and intelligent along with the high development of computer technique. Image processing and automatic welding seam tracking are two important parts of weld automation. The technique of seam tracking based on vision sensing is one of the first-class techniques in the field of seam tracking in the recent years. Because of large information, high reliability, broad applicability etc., it has a huge potential for research and application.The seam tracking system based on the USB interface technology. The system is built by USB camera. I design the program of image collection with VB6.0. On the basis of above mentioned, in-depth study on digital image processing of submerged arc welding on spiral pipesIn this paper, the image processing algorithms of submerged arc welding seam is divided into temperature field analysis and auxiliary illuminant analysis. Each situation of image processing algorithm was studied deeply. In the situation of temperature field analysis, the position of welding center will be obtained after some image processing program such as color image transfer into grayscale, grayscale projection and judge the symmetry of curve, auto-select thresholds in all image, binary image, grey grade curve. In the situation of lighting strip of auxiliary illuminant analysis, using the lighting strip of auxiliary illuminant will make the characteristic of welding seam more obvious, with the technique of opening a small window and least square algorithm, the position of welding center will be obtained after some image processing program such as color image transfer into grayscale, smoothing, edge detection, auto-select thresholds, binary image, thinning image and de noising program. this method make the precision higher, processing becomes more speed. In the situation of full-court common light analysis, the position of welding center will be obtained accurately after some image processing algorithms such as color image transfer into grayscale, auto-select thresholds in the whole image, binary image and grayscale projection. The processing time is less than 50ms.Image processing program includes smoothing, sharpening, edge detection, de noising programautomatically threshold selection of difference histogram and iteration method. Image processing program that programed by VB6.0 programming can obtain the position of welding center finally. The study is meaning to research of welding seam self-tracking system and image processing algorithms of other areas.