基于多特征提取的药品日期识别方法  

Date recognition based on multi feature extraction

作　　者：王敏 吴佳 孙硕 李晟 王康 WANG Min;WU Jia;SUN Shuo;LI Sheng;WANG Kang(School of Electronics and Information Engineering,Nanjing University of Information Science and Technology,Nanjing 210044,China;Suzhou Minjie Robot Technology Co.,Ltd.,Suzhou 215159,China)

机构地区：[1]南京信息工程大学电子与信息工程学院,江苏南京210044 [2]苏州敏捷机器人科技有限公司,江苏苏州215159

出　　处：《物联网学报》2022年第3期146-152,共7页Chinese Journal on Internet of Things

基　　金：国家自然科学基金资助项目(No.41775165,No.41775039);南京信息工程大学人才启动经费资助项目(No.2021r034)。

摘　　要：药品生产日期和有效期是衡量药品安全性和有效性的重要指标。可立袋液体药品生产日期及有效期均由0~9的印刷体数字表示,医院对药品日期的识别需要满足快速准确的要求,常规的模板匹配法、神经网络识别法计算量大、训练复杂。提出了一种基于多特征提取的药品日期识别方法,通过欧拉数、横线特征、竖线特征3个特征结合对不同数字字符进行精细化特征提取,具有计算量小、识别速度快等优点,相对于单一提取特征进行识别的方法而言,能有效地将10个不同数字区别开来,尤其适用于易混淆的数字。可为病人提供安全用药保障的同时改善医务人员工作的管理模式,并提高工作效率,进而提升医院药学服务水平。Drug production date and expiration date are important indicators to measure the safety and effectiveness of drugs.The production date and validity period of liquid drugs in vertical bags are printed with numbers of 0~9.The recognition of the drug date of the hospital needs to meet the requirements of high speed and accuracy.The conventional template matching method and neural network recognition method have large amount of calculation and complex training.A drug date recognition method was proposed based on multi feature extraction.The combination of vertical line feature and three features for fine feature extraction of different digital characters has advantages of small amount of calculation and fast recognition speed.Compared with the recognition method of single extracted feature,it can effectively distinguish 10 different numbers,especially suitable for numbers of similar shape.It can provide patients with safe drug use guarantee,improve the management mode of the medical staff,improve the work efficiency,so as to improve the level of the pharmaceutical service of the hospital.

关 键 词：有效期 字符识别 欧拉数 横线特征 竖线特征 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置]