摘要
在饮料食品行业中饮品的包装标签包含了饮品的各种信息,其好坏关系着饮品是否合格。针对近似圆柱的饮品包装标签设计了一种基于机器视觉的标签缺陷检测方法,能够多方位的采集到柱面饮品标签的图像信息。针对柱面饮品标签存在的形状畸变,提出了一种柱面反投影校正算法;将经柱面反投影后相邻标签图像进行配准,采用非线性权重融合成完整的柱面标签展开图;最后对标签展开图的颜色空间进行信息提取,采用双阈值和形态学处理来判断该饮品标签是否合格。实验结果表明该方法能较好的实现柱面饮品包装标签的缺陷检测,有望应用于饮品灌装生产线。
In the beverage and food industry,the packaging label of a beverage contains all kinds of information about the beverage,and its quality is related to whether the beverage is qualified.A defect inspection system based on machine vision was presented in this paper,which is designed for labels on cylindrical beverage bottles.This system can capture image information of cylindrical labels from multiple azimuths.Firstly,an adverse cylindrical projection algorithm is proposed to rectify the distortion of cylindrical beverage labels.Secondly,the adjacent label images are registered after the cylindrical back-projection,and the non-linear weights are used to build panorama for label inspection.Finally,the color space of the label panorama is extracted,and the double threshold and morphological processing are used to judge whether the beverage label is qualified.The experiment results show that this method can achieve the defect inspection for cylindrical beverage packaging labels,and it is expected to be applied to beverage filling production lines.
作者
梁桥康
谢兵兵
郭东妮
孙炜
王耀南
LIANG Qiaokang;XIE Bingbing;GUO Dongni;SUN Wei;WANG Yaonan(College of Electrical and Information Engineering,Hunan University,Changsha 410082,China;Hunan Key Laboratory of Intelligent Robot Technology in Electronic Manufacturing,Changsha 410082,China;National Engineering Laboratory for Robot Vision Perception and Control Technologies,Changsha 410082,China)
出处
《无人系统技术》
2020年第2期43-48,共6页
Unmanned Systems Technology
基金
国家自然科学基金(61673163)
湖南省科技创新计划项目(2017XK2102)
电子制造业智能机器人技术湖南省重点实验室开放基金(IRT2018003)。
关键词
机器视觉
柱面标签
缺陷检测
柱面反投影
颜色空间
形态学
Machine Vision
Cylinder Label
Defect Inspection
Adverse Cylinder Projection
Color Space
Morphology Process
