Research on intelligent positioning of initial welding point of mobile welding robot on curved structures

With the increasing requirements of industrial manufacturing on the positioning accuracy of welding robots, this paper proposes an intelligent positioning method for weld starting point of curved structures based on structured light depth camera, and constructs a complete technical route through the steps of point cloud preprocessing, surface smoothing, weld positioning and starting point determination. The experimental results show that the method can stably identify the weld seam and accurately locate the starting point under the conditions of complex background and multi-pose change. Compared with the traditional manual marking method, the starting point positioning accuracy is improved by 65%, and the robustness is improved by about 48%, which provides a new idea for the high-precision autonomous operation of mobile welding robots, and effectively solves the problem of insufficient adaptability of the traditional positioning method on curved structures.