Study On The Image Processing Technology Of Rapidly Identifying Characteristics During The Seam Laser Tracking

In this paper, the design of the visual sensor was proposed. The design solved the problem that the laser might damage if it was too close to the welding seam by adding a small mirror. On the one hand, it increased the distance of the laser and welding seam. On the other hand,it increased the laser light incident on a wider range.Proposed a four-step threshold segmentation algorithm and improved otsu threshold segmentation algorithm. The four-step threshold segmentation algorithm processing time was40ms. Optimize otsu threshold segmentation algorithm ultimately shortened the processing time to12ms, at the meanwhile, it improved image processing effect. Proposed improved algorithm based on SUSAN edge detection algorithm, which extracted the better edge positioning, the thinner welding ling which was only a single pixel, and more smooth edge. Finally, the entire seam image processing speed was to achieve80ms, which met the requirements of real-time tracking.Developed the laser seam tracking software under VC++6.0development platform. The software included a command-control module, a real-time observation module, a manual calibration module, a material selection module, a parameter adjustment module these five functional modules to achieve the laser real-time tracking of the welding seam.Proposed a rapid identification method for typical groove. Meanwhile analyzed the characteristics of the groove, and then propsed two optimization slope method to extract the feature points of the welding seam, which were the reciprocal of the slope method and dynamic slope method. The latter could also effectively identify U-shaped welding seam. Proposed maximum distance method to extract the feature points of the welding seam for all kinds of groove forms. Which was simple, fast and met the requirements of real-time tracking.Real-time tracking test was carried out by using size of240×120×6mm3of3003aluminum alloy for the slash and curve seam.The test was designed for typical groove such as Ⅰ,Ⅴ.half-V, U-shaped groove. As a result, when the welding speed was in60-80cm/min and the slope of joints were in the10°～40°, the slash tracking deviation was less than0.5mm, and the S-shaped,radius of curvature of50mm curve tracking deviation was less than lmm, which met the accuracy requirements of real-time seam tracking.