The Research Of Seam Tracking Experimental System Applied To Underwater Hyperbaric Welding

With the increase of submarine pipelines laying, underwater welding has been considered by increasing people. Underwater hyperbaric welding is an important method in the underwater welding field, and realizing the automation of underwanter hyperbaric welding is critical to the development of underwater welding. The technology of seam tracking is an important part of underwater hyperbaric welding.Based on the in-depth investigation of the latest progress of seam tracking technologies and the influence of hyperbaric environment upon the seam tracking system, a set of seam tracking experimental system has been established. In the system, a technology of passive vision is adopted, and a filter with an appropriate bandwidth is used to acquisition high quality seam images in hyperbaric welding environment. The latest developments in DSP technology applied to the system's control unit can ensure the system off-line mode and a real-time processing ability. Seam images are processed by DM642 using corresponding seam recognition algorithm, and the deviation information is calculated by the algorithm. The information is translated into two PWM signals with different duty cycle by real-time control algorithm in DSP2812, and the PWM signals are used to control motor's moving direction and moving speed.A feasible seam recognition algorithm is introduced on the basis of the research and analysis of image processing algorithms applied to passive vision technology and DSP technology. Respectively, windowing technology, median filtering, Sobel edge detection, binarization, edge search of a particular region are used to the seam recognition algorithm, which ensure the algorithm's real-time ability. To enhance the algorithm's anti-interference ability and make the algorithm robust, the seam midline data calculated by the algorithm is filtered using median filtering algorithm. The points of seam midline in the region at the top of tungsten pole and the points of the tungsten pole's midline are selected to be used to calculate the deviation information, which can eliminate the system's precursor error. In order to prevent the algorithm failure and stop automatically when the position of the tungsten pole is adjusted beyond the range of seam area, some control parameters are added to the algorithm.Curved and linear seams are tracked under the simulated and welding conditions respectively using the system, and the system's intelligentized ability is also validated in experiments. The experimental results show that the system can meet the requirements of underwater hyperbaric welding.