摘要
针对传统依靠人工目视进行法兰表面缺陷检测存在效率低下、受人工经验和主观因素的影响大、准确率低等问题,提出一种基于连通域分析的法兰表面缺陷视觉检测方法。首先,根据法兰图像灰度直方图特点,使用大津法(Otsu)和连通域分析相结合的方法实现图像分割。其次分析法兰表面的缺陷区域与正常区域的灰度差异、缺陷形态及位置,按缺陷类型将法兰表面划分为两个待检测区域,并分别设计算法进行缺陷检测。实验结果表明:本研究提出的法兰表面缺陷检测方法稳定性好,能够准确、快速地识别法兰表面缺陷区域,检准率可高达97.7%。
In view of the problems of low efficiency,great influence of manual experience and subjective factors and low accuracy in the traditional visual inspection of flange surface defects,a visual inspection method for flange surface defects based on Bolb analysis was proposed.Firstly,according to the characteristics of the gray histogram of the flange image,Use Otsu method(Otsu)and Bolb analysis method to achieve image segmentation.Secondly,analyze the gray difference,defect shape and posi⁃tion between the defect area on the flange surface and the normal area,divide the flange surface into two areas to be inspected ac⁃cording to the defect type,and design algorithms for defect inspection respectively.The experimental results show that the proposed method has good stability and can accurately and quickly identify the flange surface defect area,and the accuracy rate can reach 97.7%.
作者
马光岩
曾劲松
刘国宁
焦博
MA Guang-yan;ZENG Jin-song;LIU Guo-ning;JIAO Bo(School of Mechanical and Power Engineering Zhengzhou University,He’nan Zhengzhou 450001,China)
出处
《机械设计与制造》
北大核心
2024年第9期233-237,共5页
Machinery Design & Manufacture
基金
河南省产学研重点支持项目(172107000008)。
关键词
法兰
缺陷检测
连通域分析
机器视觉
Flange
Defect Detection
Bolb Analysis
Machine Vision
作者简介
马光岩,(1997-),男,河南三门峡人,硕士研究生,主要研究方向:无损检测技术;曾劲松,(1977-),男,湖北应城人,博士研究生,副教授,硕士生导师,主要研究方向:先进制造技术,无损检测技术。
