Robust pattern recognition of three dimensional weld pool surface in gas tungsten arc welding

Measuring weld pool surfaces in real-time is the foundation for intelligent welding that emulates welders´ sensory capability. To image and measure the mirror-like weld pool surface under the strong arc radiation, a structured light laser pattern, i.e., a dot matrix, is projected onto the weld pool surface and its specular reflection is intercepted and imaged by an imaging plane placed with a distance from the arc. In this paper, a robust recognition scheme is proposed to identify the reflection pattern of the dot matrix in real-time. In particular, an adaptive segmentation algorithm is developed to identify the reflection pattern. Then a pattern recognition is proposed to determine the row and column number of each laser dot in the pattern such that the reflection pattern can be matched with the projection pattern. The identified reflection pattern can be used to reconstruct the 3D weld pool surface. Experiments with different welding conditions have been conducted, and the real-time performance of the proposed procedure, including the effectiveness, robustness and time cost, has bee verified.