Computer Engineering and Applications ›› 2020, Vol. 56 ›› Issue (4): 134-139.DOI: 10.3778/j.issn.1002-8331.1811-0005

Previous Articles     Next Articles

Research on Dot Matrix Character Recognition Based on Template Matching and Support Vector Machine

MA Ling, LUO Xiaoshu, JIANG Pinqun   

  1. College of Electronic Engineering, Guangxi Normal University, Guilin, Guangxi 541004, China
  • Online:2020-02-15 Published:2020-03-06



  1. 广西师范大学 电子工程学院,广西 桂林 541004


The dot matrix character information on food and pharmaceutical packages generally includes its date of manufacture and other important information. At present, the single dot matrix character recognition method has low accuracy. In a complex environment(including both dot matrix characters and continuous characters), the dot matrix character has a low accuracy of character localization. To this end, template matching combined with Support Vector Machine(SVM) is proposed. Firstly, dot characters are accurately located by using the discrete properties of dot matrix characters. Then two decision results are respectively obtained by gray-based template matching and feature-based template matching methods. If the judgment result is the same, the character is recognized. If the judgment result is different, the two results are sent to the SVM for recognition. Finally, the recognition result is obtained. The experimental results show that the proposed method is superior to the traditional character recognition method in terms of location accuracy and recognition accuracy. The robustness is good, and the character recognition rate reaches 96.10%.

Key words: dot matrix character, Support Vector Machine(SVM), template matching, dot matrix character recognition


食品、药品包装上的点阵字符信息一般包含生产日期和其他重要信息。针对目前单一的点阵字符识别方法准确率不高,且对点阵字符在复杂环境下(既包含点阵字符又包含连续字符)字符定位准确性低的问题,提出了一种基于模板匹配和支持向量机(Support Vector Machine,SVM)的组合点阵字符识别方法。该方法利用点阵字符的离散性质来准确定位点阵字符,然后分别通过基于灰度的模板匹配和基于特征的模板匹配方法得到两个判定结果。若判定结果相同,则识别出字符;若判定结果相异,将这两个结果送给SVM进行识别,得出识别结果。实验结果表明,该方法在点阵字符的定位准确性和识别率方面都优于传统字符识别方法,且识别鲁棒性较好,字符识别率达到96.10%。

关键词: 点阵字符, 支持向量机(SVM), 模板匹配, 点阵字符识别