Research On Automatic Tracking System Of Spiral Tube’s Submerged Arc Welding Based On CCD

In the productive process of spiral tube, its welding quality is often affected by lateral disturbance of hot rolled coil, welding heat distortion, flux leakage and human factor. To overcome the influence factors in the process of submerged arc weld spiral pipe welding and to get rid of the dependence on people in welding process, making sure that the deviation of the weld is controlled in the specified range in the case of adverse factors, the experimental study of the automatic tracking system of the spiral pipe welding is carried out.Welding image of submerged arc welding is adopted into the industrial PC through CCD visual sensor which is used as sensing component and CG300 which is used as image acquisition unit in the research. Image acquisition and recognition system is designed by using VB6.0. The deviation data which is given by image processing system is received through PLC which is used as executive controller and spiral tube’s welding torch is also controlled by PLC to move to correct the welding deviation. Those make up automatic tracking system of spiral tube’s submerged arc welding.Based on the system hardware design, this study focuses on image recognition algorithm. In order to facilitate algorithm and overcome over-reliance on actual production site, the simulative welding seam is carried out and the simulative software is written. The influence of light source, flux leakage and surface state of the work pieces on weld line’s image processing in the factory is simulated in the laboratory. Threshold algorithm of welding recognition is designed. The automatic threshold method of average threshold and bimodal law is given.Through the analysis of influence of various external disturbances, the law of the influence of the different conditions on the weld projection is found out. The recognition accuracy of the welding is raised to 1-2mm; Finally the combination of the optimized image processing algorithm is derived: Firstly the image is softened to de-noising, then sharped to increase the contrast ratio and went on image’s binarization through automatic bimodal thresholding, finally extracted the lateral deviation of weld torch and seam.The control system which is controlled by PLC conducts communication by free port communication, so the time of collecting images, identifying errors, data transmission and adjusting signal from PLC is less than 100 ms. The control system could meet the control requirements of seam tracking.