Research Of Vision Sensing Seam Tracking Method For Underwater Welding Robot

With the rapid progress of exploring the sea, underwater welding robot plays an important role in the exploration of the sea. Underwater welding seam tracking is an important part of welding robot. In order to achieve an underwater seam tracking method with laser visual sensor, a system of underwater seam tracking was built. In this system, the welding seam structure optical images can be captured by CCD with a liner laser as an active light source. Images can be relayed to computer by image captureing card, Images can be processed by the computer and the weld position can be worked out, so the welding torch can be controlled.The underwater laser vision sensing environment factors and the disturb factors (?)re analyzed in detail. On this basis, a laser welding seam image sensor suiting the underwater environment is designed and produced. In the article, the sake of sensor components choice, the optical, framework and seal design were all stated.The characteristics of underwater welding seam sensing images captured by the laser welding seam image sensor is analyzed, .In order to gain the seam position from the image, many kinds of image process methods was compared, and a better group was chosen,. That is morphology filtering, the improved OTSU image segmentation, center sample, Hough transform to extract the seam signal. In order to meet real-time requirements of the welding system, improve sampling Center and the Hough transform algorithm . the software of image processing is written with VC++, which can get the seam signal accurate.Finally, after the seam signal is gotten and the weld deviation is worked out, the control system of FUZZY-PID is designed in which the weld deviation and weld deviation vary rate (which is the seam position deviation between the adjacent images) is as inpu. on account of its shortcomings, Fuzzy controller of the scale factors are optimized based on genetic algorithm optimization, so that the output of fuzzy controller is more precise. Thus overcoming the shortcomings of the single fuzzy control , in which the fuzzy rules and membership functions is difficult to determine, and the shortcomings of welding parameters on-line measurement technology ,which is weak anti-interference ability and the adjust rate is too large to control precisely.The good results is achieved . And the control algorithm software was write with C++.