基于亮度校正与连通域分割的电池镍片缺陷检测
王语言 1梁丹 1马益平 2龙翔林 2陈贤儿 1罗钦泽1
作者信息
- 1. 宁波大学机械工程与力学学院 浙江宁波 315211
- 2. 宁波迦南智能电气股份有限公司 浙江宁波 315399
- 折叠
摘要
针对电池镍片表面特性,提出一种基于亮度校正与连通域分割的电池镍片缺陷检测方法.设计基于Mask匀光校正和阈值分割的图像预处理算法,进行亮度矫正和缺陷区域的预分割.提出一种基于改进种子填充的电池镍片表面缺陷区域分割算法,去除细小噪声,并准确提取缺陷区域.设计基于几何特征的表面缺陷分类,实现典型缺陷的准确分类.试验表明,所提出的检测方法缺陷检测准确率达到93％,平均运行时间为0.861s,能够有效检测电池镍片不同类型的表面缺陷.
Abstract
According to the surface characteristic of battery nickel sheet,a defect detection method for battery nickel sheet based on brightness correction and connectivity domain segmentation was proposed.An image preprocessing algorithm based on Mask homogenization correction and threshold segmentation was designed to carry out brightness correction and pre-segmentation of defective area.A surface defect area segmentation algorithm for battery nickel sheet based on improved Seed-Filling was proposed to remove the fine noise and accurately extract the defective area.A surface defect classification based on geometric feature was designed to achieve accurate classification of typical defect.The test result shows that the defect detection accuracy of the proposed detection method reaches 93％,and the average running time is 0.861 s,which can effectively detect different types of surface defects of battery nickel sheet.
关键词
电池/镍片/亮度校正/连通域分割/缺陷/检测Key words
Battery/Nickel Sheet/Brightness Correction/Connected Domain Segmentation/Defect/Detection引用本文复制引用
出版年
2025
NETL
NSTL
万方数据