基于分段拟合算法的V型焊缝识别  

V⁃shaped weld seam recognition based on subsection fitting algorithm

作　　者：谢昊 郭天太 王浩 XIE Hao;GUO Tiantai;WANG Hao(College of Metrology and Measurement Engineering,China Jiliang University,Hangzhou 310018,China)

机构地区：[1]中国计量大学计量测试工程学院,浙江杭州310018

出　　处：《现代电子技术》2023年第20期153-158,共6页Modern Electronics Technique

摘　　要：激光视觉引导的焊接方式在工业产品线上应用广泛,其中图像处理算法的复杂性会直接影响焊接的效率。为解决激光视觉中的V型焊缝直线识别速度问题,利用“一”字激光光源实验,在分析了不同图像预处理方法对激光条纹直线特征的提取效果上,提出一种基于斜率分析的点集分割方法。该方法能够正确区分V型焊缝不同线段上的点,结合最小二乘法进行分段拟合,准确提取焊缝的中心线特征。结果表明,该方法与Hough变换相比,可避免因误差而拟合成非连续线段,具有计算量小、反应速度快、适应性强的特点。Laser vision guided welding is widely used in industrial product lines,and the complexity of image processing algorithms can directly affect the efficiency of welding.In order to solve the problem of speed recognition of V⁃shaped weld seam line in laser vision,a point set segmentation method based on slope analysis is proposed by means of the"one"laser light source experiment and analyzing the extraction effect of different image preprocessing methods on laser stripe line features.This method can correctly distinguish the points on different line segments of the V⁃shaped weld seam,and the least square method is used to conduct the piecewise fitting,so as to accurately extract the centerline features of the weld seam.The results show that in comparison with Hough transform,this method can avoid fitting discontinuous line segments due to errors,which has the features of less computation,fast reaction speed and strong adaptability.

关 键 词：焊缝识别 分段拟合 激光视觉 图像预处理 激光条纹 直线特征提取 HOUGH变换 

分 类 号：TN249-34[电子电信—物理电子学]