基于机器视觉的食品包装袋喷码字符识别方法  

Inkjet Character Recognition of Food Packaging Bag Based on Machine Vision

作　　者：姚勇 吴田 YAO Yong;WU Tian(School of Computer Science and Engineering,Hubei University of Education,Wuhan 430205,China)

机构地区：[1]湖北第二师范学院计算机学院,武汉430205

出　　处：《湖北第二师范学院学报》2023年第8期51-57,共7页Journal of Hubei University of Education

摘　　要：针对目前食品包装袋喷码字符人工识别存在劳动强度大、效率低、自动化水平低以及字符粘连难以分割等问题,本文提出一种基于机器视觉的食品包装袋喷码字符识别方法。首先对字符位置进行定位,其中包括预处理、连通域分析和特征提取,然后对字符采用第一次连通域粗分割和第二次字符细分割的二次分割方法来提高字符分割的准确率,最后利用标准的模板匹配法识别字符。实验结果表明,所提方法可以较好地实现对喷码字符的定位识别。The present manual recognition of inkjet characters on food packaging bags has problems such as high labor intensity,low efficiency and low automation level.Therefore,an automatic recognition of packaging bag inkjet characters based on machine vision is proposed.First,the character position is located,including preprocessing,connected domain analysis and feature extraction.Then,second segmentation method of first connected domain segmentation and second character fine segmentation are used to improve the accuracy of character segmentation.Finally,the standard template matching method is used to identify characters.The experimental results show that the proposed method can better realize the location and recognition of inkjet characters,and has a certain application value.

关 键 词：喷码字符 字符定位 二次分割 字符识别 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TS21[自动化与计算机技术—计算机科学与技术]