基于机器视觉的焊缝宽度测量方法  

A Method of Weld Width Measurement Based on Machine Vision

作　　者：苗璐 景文博[2] 蔡立娟[1] 山显响 MIAO Lu1, JING Wenbo2, CAI Lijuan1, SHAN Xianxiang1(1.School of Electronics and Information Engineering, Changchun University of Science and Technology, Changchun 130022; 2.School of Optoeleclmnic Engineering, Changchun University of Science and Technology, （＇hangchun 130022)

机构地区：[1]长春理工大学电子信息工程学院,长春130022 [2]长春理工大学光电工程学院,长春130022

出　　处：《长春理工大学学报（自然科学版）》2018年第1期114-118,共5页Journal of Changchun University of Science and Technology(Natural Science Edition)

摘　　要：针对焊缝检验尺对焊缝宽度检测精度差和效率低的特点,提出一种基于机器视觉的焊缝宽度测量方法。通过测量设备获取目标图像,使用直方图均衡化和分段线性变换相结合的算法对图像进行增强。采用基于最大阈值Ostu的图像分割算法经过膨胀腐蚀等形态学的处理提取目标轮廓,再利用霍夫变换检测焊缝边缘,计算焊缝宽度。实验结果表明,该算法对焊缝的宽度检测精度可达0.01mm,效率更高。A method of measuring the width of welds based on machine vision is proposed for the characteristics of weld inspection scale and low detection accuracy. Through the measurement equipment to obtain the target image,the image is enhanced by the combination of histogram equalization and piecewise linear transformation,using the maximum threshold Ostu image segmentation algorithm~[1],swelling and other morphological processing to extract the target contour,useing the Hough transform to check the edge of the weld,and then calculating the width of weld. The experimental results show that the accuracy of the weld width detection can be reached to 0.01 mm,which made the calculation more efficient,and it provides a referential meaning for other areas of mechanical automation.

关 键 词：机器视觉 图像处理 直线检测 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置]