北京邮电大学学报

  • EI核心期刊

北京邮电大学学报 ›› 2006, Vol. 29 ›› Issue (1): 13-16.doi: 10.13190/jbupt.200601.13.zhangch

• 综述 • 上一篇    下一篇

适用于银行票据手写数字串切分的滴水算法

张 闯1, 蔺志青1, 肖 波1, 郭 军2   

  1. 1.北京邮电大学 电信工程学院,北京 100876
    2.北京邮电大学 信息工程学院,北京 100876
  • 收稿日期:2004-12-03 出版日期:2006-02-28 发布日期:2006-02-28
  • 通讯作者: 张 闯 E-mail:zcbupt@sina.com
  • 作者简介:张 闯(1975—),男,讲师,E-mail: zcbupt@sina.com.
  • 基金资助:

    教育部科学技术研究重点项目(02029)

Segmentation Algorithm for Unconstrained Handwritten Numeral Strings in Bank Check Reader System

ZHANG Chuang1, LIN Zhi-qing1, XIAO Bo1, GUO Jun2   

  1. 1.School of Telecommunication Engineering, Beijing University of Posts and Telecommunications, Beijing 100876, China
    2.School of Information Engineering, Beijing University of Posts and Telecommunications, Beijing 100876, China
  • Received:2004-12-03 Online:2006-02-28 Published:2006-02-28

摘要:

在连通域提取与属性判别的基础上,得到不固定长度的粘连字串,利用波形分析的方法对字串个数进行判断,用具有指导信息的改进滴水切分算法解决了覆盖、粘连等非约束不确定位数的手写数字串切分问题. 该切分算法在实际的银行票据自动识别系统中取得了实用化的效果.

关键词: 票据识别, 数字切分, 字串个数判断

Abstract:

How to improve the strings segmentation process was investigated. The components connected were detected based on isolated components analysis, and the length estimation of the components using the feature of lower contours was discussed, the problem of segmenting overlapping or touching characters settled by the improved “drop-falling” algorithm was also studied in details. This method has been incorporated into a complete bank check character recognition system. It shows promising results.

Key words: bank check recognition, segmentation, length estimation of digit string

中图分类号: