基于相关系数的磁墨字符信号识别算法

Recognition Algorithm of Magnetic Ink Character Signal Based on Correlation Coefficient

通过研究当前国内磁码的识别方法,提出了一种基于相关系数的磁墨字符信号快速识别算法。该算法是以支票磁码信号为识别目标,以统计分析方法,确立了检测阈值,运用Matlab软件将采集到的整张支票的磁码长信号进行一系列的预处理及信号分离,把分离出的单个磁墨字符信号与每个标准磁墨字符信号求相关,以相关系数作为待识别磁码与标准磁码的相似性度量标准,以最大相关系数所对应的标准磁墨字符符号作为最终识别结果,实现整个磁码信号的识别。对大量的实验结果进行统计分析得出,该识别算法运算速度快、实时性好,识别精度达99.9%以上。

Through the research of the current domestic magnetic code identification methods,the fast recognition algorithm of magnetic ink character signal on basis of a correlation coefficient was proposed.The algorithm regards the magnetic code signal of the check as the identification target,using the statistical analysis methods to establish the detection thresholds,and using Matlab software to have a series of pre-processing and signal separation for the collected magnetic code signal of the whole check.In correlation recognition algorithm,the maximum correlation coefficient between the isolated single magnetic ink character signal and each standard magnetic ink character signal is used to determine the recognition result.The statistical analysis from a large number of experimental results prove that the recognition algorithm has a high operation speed,a good real-time,and it＇s recognition accuracy can reach 99.9%.