多个数字仪表动态显示数字字符识别的研究  

Research on recognition of dynamic characters in multi digital instruments

作　　者：李旦[1] 隋成华[2] 唐轶峻[2] 

机构地区：[1]浙江工业大学信息工程学院,浙江杭州310032 [2]浙江工业大学理学院,浙江杭州310032

出　　处：《浙江工业大学学报》2007年第4期437-440,共4页Journal of Zhejiang University of Technology

基　　金：浙江省科技厅资助项目(2004030063)

摘　　要：在诸多像冶金、化工等的行业,都要用到多个数字仪表的实时监控.而人眼识别会产生误判,且效率不高,用多个数字仪表动态字符识别系统代替人眼将提高效率,同时识别率的提高可以解决人眼产生误判的问题.然而这一系统研制的难点在于多仪表数显字符的快速分割定位和实时识别,为了解决快速分割,提出了将特征颜色增强的方法应用于该系统.同时为满足实时识别,采用了BP神经网络法进行分类识别.实验结果表明,在小于3个数字仪表的条件下,整个识别过程可控制在40 ms以内,正确识别率达到96%以上.In many as metallurgy, the chemical industry, a number of digital real-time monitoring instrument are used. The manual method will bring the problems of inefficient and misjudge. Digital Instrumentation with multiple dynamic character recognition system will replace the human eye to enhance efficiency and resolve the issue of misjudgment. However, the two critical issues in the system developed are rapid segmentation and real-time positioning identification. To address the rapid segmentation, the feature color method is used to enhance the system. At the same time, to meet the real-time identification, the BP neural network for classification is used. Experimental results show that the entire identification process canbe controlled in less than 40 ms less than three meters and correct identification rate is over 96%.

关 键 词：多数字仪表动态字符识别 特征颜色 BP神经网络 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置]