Study And Development Of On-Line Detection And Tracking System Of Seam In Helix Welded Pipe

Seam tracking control is the key during welding process, which plays an important role in the automatic control of welding quality. Seam detection is also one of the kernel technologies of seam tracking. Many types of sensors were applied to seam detection process, which including vision sensor, arc sensor and so on. It is a trend technology that implementing seam detection using vision sensor.This paper introduced a visual servo and control system for welding seam tracking including the hardware configure, the method for welding seam recognition, the control strategy for welding seam tracking and its experimental results in practice.Filter methods of image processing and edge detection is used to accomplish the pretreatment of the image, and is found to be able to reduce the noise and detect the edge, and make the welding seam clear. After an experimental comparison of several methods of edge detection in weld seam image processing, Roberts operator is found to be more effective in systematical image edge detection.To solve the problem that the position sensing is often delayed by image processing on computer, this paper adopted a close loop control method of position tracking, in which results of position vision sensing were input as position commands, pulses from the encoder combined with motor axis were fed back as speed signal and pulses from the raster ruler with whorl axis nut were fed back as position signal. This method was able to avoid some negative effects produced by the delayed units on the stability and dynamic performance of the system.After an analysis of the sources of errors in the vision servo tracking system and the effects of different errors on tracking performance, it was found that the automatic seam tracking system designed in this paper could meet the technical requirements. Being simple and practical in structure, the system was easy to operate. It has some innovations while it has inherited the advantages of the previous ones in terms of fulfilling the requirements of both accurate edge detection and real time control. The system was highly practical.