复杂背景下的粉笔数字字符自动提取方法研究  

Automatic extraction method of chalk characters in context of complex

作　　者：吴炜[1,2] 骆剑承[1] 胡晓东[1,2] 沈占锋[1] 陈亮 

机构地区：[1]中国科学院遥感应用研究所,北京100101 [2]中国科学院研究生院,北京100049 [3]南京梅山钢铁厂自动化二室,南京210000

出　　处：《计算机应用研究》2009年第10期3963-3965,共3页Application Research of Computers

基　　金：国家科技支撑计划重大资助项目(2006BAJ02A01);国家"863"计划资助项目(2007AA12Z141)

摘　　要：根据粉笔字符背景复杂、断笔、残笔、手写字符粘连、重叠等特点,提出了一种复杂背景下粉笔数字字符自动提取方法。通过数字区域定位,获取局部图像并进行二值化;通过数学形态学处理,解决二值化结果中的断笔、残笔等问题;通过改进滴水算法,对粘连字符进行切分,获取单个字符。最后,以板坯号识别为例,开发了板坯号自动识别系统,并成功地应用在板坯的生产过程中,证明了其有效性。The extraction of the slab ID was taken as example, this paper proposed an automatic exaction method for chalk characters, through the location the region of the chalk area, obtained and binaried the local image, processed the result by mathematical morphology method and solved the problem of non-consecutive stroke. Through the improvement of the ＂drop falling＂ algorithm, solved the problem of overlapping and touching characters. Last, realized the algorithm by programming and developed a slab ID system, used it in the product line successfully and proved efficient.

关 键 词：粉笔字符 自动提取 区域定位 图像二值化 字符切分 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]