运用灰色关联度的仪表数字字符识别-《集美大学学报（自然科学版）》

文章信息/Info

Author(s):: CHEN Hongyu1; LI Tiejun2; (1.School of Marine Engineering，Jimei University，Xiamen 361021,China;2.School of Information Engineering，Jimei University,Xiamen 361021,China）

Keywords:: grey correlation degree; hardness feature; meter digital; digital identification

摘要:: 针对仪表数字字符图像识别中存在角度倾斜和非整字的不确定因素，提出一种基于灰色关联度的仪表数字字符识别方法。通过构建硬度特征序列，将字符特征信息中的不确定信息当作“灰信息”，利用灰色理论中的灰色关联分析法，分析待测字符特征序列和训练集样本的特征序列之间的相关程度，将最大关联度对应的样本字符作为识读结果。实验结果表明，该算法简单有效且识别率高。

Abstract:: Aimed at the uncertain factors of angle tilt and half-word in direct reading and digit recognition intelligent meters,a method of instrument digital character identification based on grey correlation degree is proposed.By constructing hardness feature sequences,the uncertain information in the character feature is regarded as “grey information”;by analyzing the correlation between the character feature sequence to be tested and the feature sequence of training samples,the maximum correlation degree is obtained as the sample recognized result.The experimental results show that the method is simple and efficient with high accuracies.