手写粉笔数字自动识别方法研究  

Automatic recognition method of chalk handwritten numerals

作　　者：胡晓东[1] 骆剑承[1] 吴炜[1] 程熙[1] 沈占锋[1] 贾银芳[2] 

机构地区：[1]中国科学院遥感应用研究所,北京100101 [2]上海梅山钢铁股份有限公司,南京210039

出　　处：《计算机工程与应用》2011年第2期182-184,248,共4页Computer Engineering and Applications

基　　金：国家高技术研究发展计划(863)(No.2007AA12Z181;No.2007AA12Z141);国家自然科学基金(No.40871203;No.40601057)~~

摘　　要：手写粉笔字在生产生活中应用广泛,针对手写粉笔字易与背景混淆的特点,提出了手写粉笔数字自动识别方法,分别通过粉笔字普适性分割算法和基于SVM的手写数字分层分类自动识别技术解决了分割和识别两个关键问题,并结合钢材板坯号自动识别系统的应用进行方法的实验,达到了较高的识别率,具有重大的应用价值和推广意义。Chalk handwriting is widely used in the production or usual life.After analyzing chalk handwriting features,the automatic recognition method of chalk handwritten numerals is proposed.Two main issues of segmentation and recognition are solved through the universal chalk handwritten segmentation algorithm and the hierarchical classification automatic recognition technique based on SVM.The experiment is executed during the application of the automatic recognition system of steel slab numbers,which proves that the method can achieve high recognition ratio.This method can reach the requirements of application,and has high value of application and significance of promotion.

关 键 词：粉笔字分割 手写数字 特征筛选 分层分类 自动识别 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]