THE STUDY ABOUT THE ALGORITHM OF THE DETECTING FEATURE ON WELD IMAGES DURING LASER WELDING PROCESSES

摘要

In laser welding important components, the precise seam tracking usually need to be used to assure the quality of final weld, because of small laser beam focus diameter and high welding speed. In the seam tracking technology, one of the key problems is the extracting of the weld track feature at high speed and accuracy. In this paper, based on wavelet transform algorithm, a novel edge detection method was proposed to be used on pattern recognition in weld seam tracking applications. This wavelet edge detection method can offer an effective solution to the contradiction between measuring precision and dealing speed during in-real time seam tracking. So it can make the precise and high speed seam tracking to come true during laser welding. Many experiments results show that by using this method more integrity edges and more details can be abstained easily, and the more precise weld position can be acquired. Finally, the good quality welds have been completed in many Ti-alloy components.