【摘要】 针对多种类结构件角焊缝特征曲面变化小，设计了一种基于线激光的角焊缝检测方法。利用机器人搭载线激光相机扫描焊缝获取点云，预处理后采用确定随机种子的RANSAC算法分割，基于最小二乘算法拟合母材平面与焊缝二次曲面，提出一种基于临域迭代均值法求取焊趾方法，计算角焊缝特征。以角焊缝工件作为实验对象，与传统结构光角焊缝检测算法对比，结果表明，该算法可以将检测进度误差控制在0.1mm以内；且重复检测偏差在±0.03mm,具有较好的准确性与稳定性。更多还原

【Abstract】 For various types of structural components with minimal changes in the characteristics of fillet welds, a method for detecting fillet welds based on line laser is designed. Utilizing a robot-mounted line laser camera to scan the welds and obtain a point cloud, the RANSAC algorithm with determined random seeds is employed for segmentation after preprocessing. The mother material plane and fillet weld’s quadratic surface are fitted using the least squares algorithm. A method based on local iterative mean is proposed to calculate the weld toe, determining the characteristics of the fillet weld. Using fillet weld workpieces as experimental objects, compared with the traditional structured light fillet weld detection algorithm, indicates that the proposed algorithm can control the detection progress error within 0.1mm. The repeated detection deviation is also within ±0.03mm, demonstrating good accuracy and stability.更多还原