摘要
针对角接、搭接、V型坡口对接和窄对接等四种典型焊缝特征的提取问题,搭建了一套基于激光视觉的焊缝图像采集系统,通过图像预处理和焊縫激光条纹中心线提取,建立了基于分段线性拟合的焊缝特征提取方法。该方法通过计算斜率划分中心线像素坐标点集,然后采用最小二乘法对各分段点集进行直线拟合来实现焊缝特征点的有效判别。结果表明:特征点的平均提取时间为9 ms,平均提取误差为0.035 mm,为基于线激光视觉图像的焊缝跟踪系统提供判据。
Aiming at weld feature extraction of the corner joints,the lap joints,the V-groove butt joints and the narrow butt joints,a set of weld image acquisition system based on laser vision is developed.Through image preprocessing and weld laser strip centerline extraction,a weld feature extraction method based on piecewise linear fiting is established.The method divides the center line pixel coordinate point set by caleulating the slope,and then the least square method is used to fit the line of each segment point set to realize the effective discrimination of weld feature points.The results show that the average extraction time of feature points is 9 ms,and the average extraction error is 0.035 mm,which provides a criterion for seam tracking system based on line laser vision image.
作者
王濠
李亮峰
张延松
WANG Hao;LI Liangfeng;ZHANG Yansong(Shanghai Key Laboratory of Digital Manufacture for Thin-walled Structures,Shanghai Jiao Tong University,Shanghai 200240,China)
出处
《机械设计与研究》
CSCD
北大核心
2021年第6期122-127,共6页
Machine Design And Research
基金
工信部高技术船舶科研项目(MC-201906-Z01)。
关键词
图像预处理
中心线提取
分段线性拟合
特征点提取
焊缝跟踪
image preprocessing
centerline extraction
piecewise linear ftting
feature point extraction
seam tracking
