首页 | 本学科首页   官方微博 | 高级检索  
     

基于机器视觉的奶粉罐体商标区域定位方法
引用本文:许敏,马钺,陈帅,徐首帅.基于机器视觉的奶粉罐体商标区域定位方法[J].包装工程,2018,39(21):27-31.
作者姓名:许敏  马钺  陈帅  徐首帅
作者单位:中国科学院沈阳自动化研究所,沈阳110016;中国科学院机器人与智能制造创新研究院,沈阳110016;中国科学院沈阳自动化研究所,沈阳110016;中国科学院机器人与智能制造创新研究院,沈阳110016;中国科学院沈阳自动化研究所,沈阳110016;中国科学院机器人与智能制造创新研究院,沈阳110016;中国科学院沈阳自动化研究所,沈阳110016;中国科学院机器人与智能制造创新研究院,沈阳110016
基金项目:广州市科技计划(201604016033)
摘    要:目的 针对奶粉包装中高盖锁扣与罐体商标图案之间位置的不确定会影响美观和品牌关注度的问题,提出一种罐体商标图案定位方法。方法 首先由透视投影变换模型获取罐体边缘上一对边缘点的摄像机坐标系下的坐标,并结合视线交叉原理,确定罐体截面圆在摄像机坐标系下的方程。其次通过模板匹配算法提取罐体上显著图案中心坐标,并获得该中心坐标与摄像机光心连线所成直线的方程。然后联立直线与截面圆的方程求交点,该交点表示显著图案中心点在截面圆上的位置。由于显著图案与商标图案相对位置已知,因此由显著图案中心坐标就可以间接求得商标图案中心坐标。结果 该方法可完成罐体商标图案的快速测定,且测量精度达到±1°。结论 该方法能够快速、准确地实现罐体商标区域的在线测量。

关 键 词:商标图案定位  模板匹配  透视投影变换
收稿时间:2018/3/9 0:00:00
修稿时间:2018/11/10 0:00:00

Location Method of the Trademark Area of Milk Powder Cans Based on Machine Vision
XU Min,MA Yue,CHEN Shuai and XU Shou-shuai.Location Method of the Trademark Area of Milk Powder Cans Based on Machine Vision[J].Packaging Engineering,2018,39(21):27-31.
Authors:XU Min  MA Yue  CHEN Shuai and XU Shou-shuai
Affiliation:1.Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang 110016, China;2.Institute for Robotics and Intelligent Manufacturing, Chinese Academy of Sciences, Shenyang 110016, China,1.Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang 110016, China;2.Institute for Robotics and Intelligent Manufacturing, Chinese Academy of Sciences, Shenyang 110016, China,1.Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang 110016, China;2.Institute for Robotics and Intelligent Manufacturing, Chinese Academy of Sciences, Shenyang 110016, China and 1.Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang 110016, China;2.Institute for Robotics and Intelligent Manufacturing, Chinese Academy of Sciences, Shenyang 110016, China
Abstract:The work aims to propose a method to position the pictorial trademark of milk powder cans regarding the problem that the uncertainty between the high cover lock of the milk powder package and the pictorial trademark of the can will affect the product aesthetics and brand awareness. First of all, the coordinates of a pair of edge points on the can edge in camera coordinate system were obtained through the perspective projection transformation model, and the equation of sectional circle of the cans was determined in camera coordinate system combined with the principle of lines-of-sight intersection. After that, the central coordinate of a significant pattern on the can was extracted by a temple matching algorithm. Hence, a line of sight between the central coordinate and the camera optical center was obtained. Then, the cross point was calculated based on the simultaneous equation of the straight line and sectional circle. The cross point indicated the position of center point of significant pattern on the sectional circle. As the relative position between the significant pattern and the pictorial trademark was known, the central coordinate of pictorial trademark could be indirectly obtained based on the central coordinate of the significant pattern. The proposed method was able to quickly measure the pictorial trademark of the cans with an accuracy up to ±1°. The proposed method can measure the trademark area of the cans online quickly and accurately.
Keywords:pictorial trademark positioning  temple matching  perspective projection transformation
本文献已被 万方数据 等数据库收录!
点击此处可从《包装工程》浏览原始摘要信息
点击此处可从《包装工程》下载全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号