金刚石线表面缺陷视觉检测方法
Visual inspection method for surface defects of diamond wire
戴福全 1池生烨1
作者信息
- 1. 福建理工大学 机械与汽车工程学院,福建 福州 350118
- 折叠
摘要
针对当前金刚石线表面质量检测算法存在因颗粒堆积、遮挡、粘连等问题导致准确率降低的情况,提出一种基于金刚石线的形貌特征和几何特征的质量检测方法.通过两台工业相机以背光照明的方式获取金刚石线图像,并应用图像处理技术进行裁剪、校正、增强、二值化等操作以准确提取金刚石线区域和特征信息,根据相关的特征信息完成对偏厚、偏薄、线头、剥离、杂质、分布、无缺陷等金刚石线图像的检测实验.实验结果表明,提出的方法能够准确、高效地提取金刚石线的各类缺陷,准确率达 98.8%,能够满足企业实际检测需求.
Abstract
In view of the fact that the accuracy of the current diamond wire surface quality detection algorithm is reduced due to the problems of particle accumulation,occlusion and adhesion,a quality detection method based on the morphological and geometric features of diamond wire was proposed.The diamond wire image is acquired by two industrial cameras in the form of backlight illumination,and the image processing technology is applied for cropping,correcting,enhancing,binarization and other operations to accurately capture the dia-mond wire area.Then,the diamond wire area is used for feature extraction,and the detection experiment of the diameter oversize,diameter undersize,wire end,stripping impurity,distribution,and no defects of the diamond wire images is completed according to the relevant feature information.The experimental results show that the proposed method can accurately and efficiently extract various defects of diamond wires,and the de-tection accuracy is 98.8%,which can meet the actual detection needs of enterprises.
关键词
机器视觉/金刚石线/质量检测/特征提取/分类识别Key words
machine vision/diamond wire/quality inspection/feature extraction/classification引用本文复制引用
基金项目
福建省中央引导地方科技发展资金项目(2022)(2022L3014)
出版年
2024
NETL
NSTL
万方数据