Application of an enhanced RANSAC method for accurate weld seam point cloud

Welding technology is a core process in modern manufacturing, and its intelligent upgrading is crucial for enhancing quality and efficiency. To address issues such as noisy point clouds and inaccurate feature extraction in the automated welding of circular plug welds, this paper proposes a robotic welding system guided by 3D point cloud data. By constructing a composite point cloud processing workflow using normal filtering, Euclidean clustering, and an enhanced RANSAC (Random Sample Consensus) algorithm, the system effectively resolves the challenge of extracting circular hole features under complex working conditions with noise exceeding 20%. This workflow enables a center positioning error of ≤ 0.05mm and a radius measurement error of ≤ 0.08 mm. Compared with conventional 2D vision methods (error of ± 1.5mm), the proposed approach significantly improves precision; welding efficiency increases by approximately 30% over manual welding, with positioning repeatability of ± 0.1mm. Moreover, uniform weld bead formation has been achieved on both flat plates and circular ring workpieces. This study efficiently addresses the precise extraction of circular hole welds from high-noise point clouds.