【作者】 吴亮； 陈宗海； 王智灵； 徐萧萧；

【Author】 Wu Liang Chen Zong-hai Wang Zhi-ling Xu Xiao-xiao (Department of Automation,University of Science and Technology of China,Hefei,Anhui,230027)

【机构】 中国科学技术大学自动化系；

【摘要】 字符识别是车牌识别过程中的关键技术,一般的方法需要庞大的样本数据和复杂计算,本文通过分析标准字符的结构知识,针对不同视角的车牌字符识别问题,提出了一种将字符模式提取与特征编码相结合的算法．该算法根据字符自身形态变化的规律,建立特征编码表,通过两次扫描并查表实现精确识别,克服了从不同角度摄取的车牌图像需要不同模板的问题,减少了计算量,提高了识别效率,在正确分割字符的前提下,识别率达到95%以上．更多还原

【Abstract】 Character recognition is an essential technology in license plate recognition,which involves numerous data and complex operation.By analyzing standard characters,this article proposed a novel algorithm in character recognition,which consists of pattern extraction and feature encoding.In this algorithm a feature table was firstly constructed for the characters.After image scanning and pattern matching,all characters could be recognized in correct.This method solves the problem of processing photographs captured from different angles with the same template,which is more fast and efficient.With correct segmentation,the precision could be high as95%.更多还原