Research On Vision-Based Tracking System For Pipeline Welding Robot

In northwest China there are abundant resources of oil and natural gas, but tremendous market requirement are there in southeast China. So the enormous separation between product and sale promotes of pipeline transportation industry. Recently, the pipeline construction technology relies on manual welding or orbited-welding device manipulated by workman, which enormously restricts the development of pipeline transportation industry. Therefore, development of intelligent pipeline welding robot will bring immense economic and society benefit.Pipeline welding robot belongs to the research field of special type robotics. Aimed at automatic girth welding of the large diameter pipelines on condition of field operations, researches are lucubrated on modeling and parameters selection of structured light sensor, processing methods of seam images, seam tracking of mobile welding robot. A seam tracking system for girth-welding robot based on vision sensing is developed.The model of a vision sensor based on structured light is educed with perspective imaging and coordinate transform principle. Using luminosity theory and considering the function of optics imaging system and light filter, the model of sensing image influenced by arc noise is established, with the distance between the arc point and the structured light stripe on the work-piece and the angle between the axis of CCD camera and the plane of structured light. With the model, approaches are derived to decrease arc disturbance and improve quality of sensing images, guiding to optimize the structure parameters of structured light sensors.Adaptive thresholding segmentation method used Maximal square difference within selected section based on priori information is developed. A compounding including this developed method, one dimension discrete LoG (Laplacian of Gaussian) filter, neighborhood averaging filter, labeling technique for small dollops removal is to be used to pre-process seam images. Then the skeleton of seam is extracted using morphologic thinning. As a result, the characteristic information is extracted rapidly and stably. Experiments show that by using this compounding processing, the seam tracking system acquires powerful anti-jamming ability. Within 0.08s, a 400×300 picxels image can be processed, so it can meet the requirement of field welding.Aimed at the pipeline welding robot, the controller for girth locomotion is designed using Lyapunov stability theory. A seam tracking arithmetic based on cylinder coordinate is developed aimed at pipeline girth welding robot, and compensate value is calculated when axis of welding robot departs from the pipeline axis.A prototype of pipeline welding robot with welding torch positioned at one side of welding robot is developed. Attaching on one of welded pipelines by permanent magnet wheels, it avoids effect of the staggered butt joint on precision of seam tracking, compared with the structure straddling on both side of welded pipeline. For an extended work, it can be used on welding of butt joint of two non-coaxial pipelines.A structured light sensor and the control system using LabVIEW with image picking, inclinometer information picking and servo motors control are taken in practice and the soft program is worked out.Seam tracking experiment shows the seam tracking system is with high reliability, and meets the requirement of pipeline girth welding.