基于机器视觉的镍板表面气孔分割算法研究
Research on stoma segmentation algorithm of nickel plate surface based on machine vision
朱正云 1许平 1陆迎东 1王磊 2高双飞2
作者信息
- 1. 昆明理工大学机电工程学院,云南昆明 650500
- 2. 金川集团股份有限公司,甘肃金昌 737100
- 折叠
摘要
针对金属镍板表面气孔缺陷检测存在分割目标堆叠和分割线偏移等问题,提出了一种基于机器视觉的镍板表面气孔分割方法,以质心标记计算距离图.该算法采用工业相机构建机器视觉系统,获取镍板表面图像.通过分段线性灰度变换抑制不感兴趣区域,其次对G通道分量进行自适应阈值分割、形态学处理等提取气孔区域,然后图像作差提取出极小值标记,以极小值标记的质心计算距离图并分水岭分割.实验结果表明,该算法气孔正确分割率达94.7%,同时能够获得更准确的气孔分割效果.
Abstract
Aiming at the problems such as segmentation target stacking and secant line deviation in detection of metal nickel plate surface stomatal defects,a method of stomatal segmentation on nickel plate surface based on machine vision is proposed,which use the centroid marker to calculate the distance map.This algorithm uses industrial camera to construct machine vision system to obtain the surface image of the nickel plate.The regions of no interest are suppressed by piecewise linear gray transformation,and the stomatal region is then obtained by adaptive threshold segmentation and morphological processing of G-channel components.The minimum marks are then extracted by image difference,the distance map is calculated with the centroid by the minimum marks and segmented by the watershed.Experimental results show that the correct stomatal segmentation rate of this algorithm is 94.7%,and it can achieve more accurate stomatal segmentation effect.
关键词
镍板/图像分割/距离变换/分水岭算法Key words
nickel plate/image segmentation/distance transform/watershed algorithm引用本文复制引用
出版年
2024
NETL
NSTL
万方数据