基于线结构光的焊缝光条亚像素中心提取方法  

Welding Seam Light Strip Sub-Pixel Center Extraction Method Based on Line Structured Light

作　　者：杨宏韬[1] 李浩 白昊天 程云龙 李秀兰[1] 张鹏[2] YANG Hongtao;LI Hao;BAI Haotian;CHENG Yunlong;LI Xiulan;ZHANG Peng(School of Electrical and Electronic Engineering,Changchun University of Technology,Changchun 130012,China;Tianjin Branch,China National Petroleum Corporation Pipeline Engineering Co.,Ltd.,Tianjin 300457,China)

机构地区：[1]长春工业大学电气与电子工程学院,长春130012 [2]中国石油天然气管道工程有限公司天津分公司,天津300457

出　　处：《组合机床与自动化加工技术》2025年第4期66-70,77,共6页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：吉林省科技发展计划项目(20240404055ZP);吉林省教育厅科学技术研究项目(JJKH20240856KJ)。

摘　　要：随着工业的迅速发展,线结构光在焊缝检测中的应用越发广泛,针对线结构光测量中焊缝特征区域曲率变化大引起的光条灰度分布不均匀,从而导致提取中心存在一定偏差的问题,提出一种基于线结构光的焊缝光条亚像素中心提取方法。该方法主要包括初始提取和精细提取两个阶段,在图像预处理的基础上,采用骨架细化法对光条中心初始提取;在精细提取阶段,利用Hessian矩阵求解其绝对值最大特征值对应的特征向量确定初始中心法向,最后在初始中心法向上引入灰度质心的思想,对法向像素集灰度进行二次平方加权计算,获取准确的亚像素中心。实验表明,所提方法解决了光条曲率变化较大的干扰问题,提取的平均偏差为0.224 7 pixel,满足工业测量的要求。With the rapid development of industry,line structured light is increasingly used in weld inspection.In order to solve the problem of uneven grayscale distribution of light strips caused by large curvature changes in the weld characteristic area during line structured light measurement,which leads to a certain deviation in the extraction center,a method for extracting the sub-pixel center of welding seam light strips based on line structured light is proposed.This method mainly includes two stages:initial extraction and fine extraction.Based on the image preprocessing,the skeleton thinning method is used to initially extract the center of the light strip;in the fine extraction stage,the Hessian matrix is used to solve the corresponding maximum eigenvalue of its absolute value.The feature vector determines the initial center normal.Finally,the idea of gray centroid is introduced in the initial center normal,and a quadratic square weighted calculation is performed on the gray of the normal pixel set to obtain the accurate sub-pixel center.Experiments show that the proposed method solves the interference problem caused by large changes in the curvature of the light strip,and the average deviation extracted is 0.2247 pixel,which meets the requirements of industrial measurement.

关 键 词：线结构光 焊缝中心 骨架细化 HESSIAN矩阵 亚像素 

分 类 号：TH16[机械工程—机械制造及自动化] TG66[金属学及工艺—金属切削加工及机床]