计算机应用与软件Issue(9)：188-191,4.DOI:10.3969/j.issn.1000-386x.2015.09.045
用于叠层纸张图像分析计数的快速直线增长算法
A FAST STRAIGHT-LINE GROWTH ALGORITHM FOR ANALYSING AND COUNTING STACKED-PAPER IMAGES
摘要
Abstract
Quantity measurement of stacked-sheet is an essential technical step in packaging and printing production process,and its counting accuracy has a direct impact on the economic efficiency of related companies.With its merits of noncontact,non-destructivity and real-time measurement,the counting method using machine vision imaging has been widely applied to quality control in high-end printing products.Aiming at the end face fringe detection problem in stacked-sheet images,it is able to improve the precision and completeness of fringe identification by improving the line-segment extraction algorithm which is based on guiding the growth along horizontal line direction,so as to realise the high accuracy measurement of stack quantity.Our work mainly consists of three parts:1 )Introducing the unidirectional gradient operator to eliminate the multiple responses on a single fringe;2)Combining the gradient magnitude and horizontal line direction, and improving the growth of line supported regions;3)Using connected element analysis to make up the anomaly of fringe’s partial rupture. At last,the performance of our algorithm is verified in experiments of actual measurements using a large number of various printed-papers. Results show that the long-term measurement error is less than 0.75‰ and it is sufficient to meet the requirement of factory applications.关键词
机器视觉/纸张计数/直线分割/LSDKey words
Machine vision/Sheet counting/Line segmentation/LSD分类
信息技术与安全科学引用本文复制引用
杨硕,彭双,肖昌炎..用于叠层纸张图像分析计数的快速直线增长算法[J].计算机应用与软件,2015,(9):188-191,4.基金项目
国家自然科学基金项目（61172160）；湖南省自然科学常德联合基金项目（12JJ9019）。 （）
