计算机应用与软件2013，Vol.30Issue(3)：108-110,117,4.DOI:10.3969/j.issn.1000-386x.2013.03.029
一种低图像质量车辆牌照的字符分割方法
A CHARACTER SEGMENTATION METHOD FOR DEGRADED VEHICLE LICENSE PLATE
摘要
Abstract
This paper proposes a novel character segmentation method for degraded license plate which are usually caused by uneven light, dirty surface or something else. Firstly, connected component analysis with OCR and plate correction will be used to obtain high confidence character regions. Then according to license plate format prior knowledge, an algorithm based on character gap analysis calculates the positions of rest characters and finally complete the whole segmentation process. This method could effectively avoid the stain, tilt and plate frame' s bad effect. The experimental results on 2 different data set prove that it is a practical and robust segmentation approach.关键词
低质量车牌/字符分割/连通区域分析/字符识别Key words
Degraded license plate / Character segmentation /Connected component analysis / OCR分类
信息技术与安全科学引用本文复制引用
王琰滨,蒋龙泉,冯瑞..一种低图像质量车辆牌照的字符分割方法[J].计算机应用与软件,2013,30(3):108-110,117,4.基金项目
上海市科委科技创新行动计划项目(10511500703) （10511500703）
上海市教委科研创新项目(产学研类)(11CXY01) （产学研类）
宝山区科委产学研合作项目(CXY-2010-35). （CXY-2010-35）
