摘要
目前我国药包生产企业普遍采用质检员抽检的方式对药品进行质量检测,人工抽检存在漏检、误检、效率低下等诸多弊端。本文提出采用基于机器视觉的检测技术对安瓿瓶的尺寸和缺陷进行在线全检,实验结果表明,该方法明显提高了检测效率和检测质量。
At present,pharmaceutical package manufacturers generally adopt the method of random inspection by quality inspectors to conduct quality testing on drugs in China.Manual sampling has many disadvantages such as missed detection,false detection,and low efficiency.This paper proposes to use machine vision-based inspection technology to perform full online inspection of the ampoule size and defects.The experimental results show that this method significantly improves the inspection efficiency and inspection quality.
作者
任亚恒
高晓芝
程煜
郝存明
陈宏彩
REN Ya-heng;GAO Xiao-zhi;CHENG Yu;HAO Cun-ming;CHEN Hong-cai(Institute of Applied Mathematics,Hebei Academy of Sciences,Shijiazhuang Hebei 050000,China;School of Electrical Engineering,Hebei University of Technology,Shijiazhuang Hebei 050018,China)
出处
《河北省科学院学报》
CAS
2020年第2期20-23,共4页
Journal of The Hebei Academy of Sciences
关键词
机器视觉
亚像素
神经网络
特征分析
Machine vision
Sub-pixel
Neural network
Feature analysis