摘要
针对膜式燃气表表观缺陷用普通拍照难以高精、高效检测的难题,提出了一种基于机器视觉的燃气表表观缺陷自动检测方法。依据均值滤波算法去除图像中噪声后,基于Otsu自适应分割方法精准分割缺陷与背景区域,在此基础上,采用连通域算法对缺陷轮廓进行提取。实验结果表明,该方法针对划痕、孔洞、表皮剥落等缺陷的检测精度达100%,检测速度为0.42f/s以内,可以有效解决当前燃气表缺陷检测技术效率低、成本高的难题。
Aiming at the difficulty of high-precision and high-efficiency detection of the apparent defects of the diaphragm gas meter by ordinary photography,an automatic detection method for apparent defects of gas meter based on machine vision was proposed.After removing the noise in the image according to the mean filter algorithm,the defect and the background area are accurately segmented with the Otsu adaptive segmentation method.On this basis,the connected domain algorithm is used to extract the defect contour.The experimental result shows that the detection accuracy of this method for defects such as scratches,holes,and skin peeling can reach 100%,and the detection speed is within 0.42 f/s,which can effectively solve the problems of low efficiency and high cost of the current gas meter defect detection technology.
作者
卢其伦
闫奕樸
张圆明
陈敏怡
汤梓玥
LU Qilun;YAN Yipu;ZHANG Yuanming;CHEN Minyi;TANG Ziyue
出处
《计量与测试技术》
2022年第7期54-56,共3页
Metrology & Measurement Technique
基金
广东省市场监督管理局科技项目(项目编号:2021CZ07)
广州市市场监督管理局科技项目(项目编号:2020KJ35)。
关键词
膜式燃气表
机器视觉
均值滤波
算法
diaphragm gas meters
machine vision
mean filter
algorithm
