基于图像处理方法的纸币清分系统的研究与实现  

On Banknote Sorting System Design Based on Image Processing Technology

作　　者：柳春华[1] 何勇[2] 

机构地区：[1]信阳师范学院计算机与信息技术学院,河南信阳464000 [2]信阳农业学院计算机科学系,河南信阳464000

出　　处：《西南师范大学学报（自然科学版）》2014年第5期81-87,共7页Journal of Southwest China Normal University(Natural Science Edition)

摘　　要：纸币自动清分可以有效提高银行的工作效率,基于图像处理技术的纸币自动清分系统的识别准确率对于纸币自动清分系统非常关键.基于机器视觉知识与模式识别的理论,并结合纸质人民币的图像特点,提出了纸质人民币序列号图像识别处理算法、"奖惩"机制下的线性传感算法和序列号区域特征提取方法.对纸币图像进行倾斜校正、局部特点提取、灰度增强和图像分块等处理后,将上述方法运用于采用CIS传感器的纸币图像处理系统中.实验证明,该方法具有较高的识别准确率,取得了理想的纸币自动清分效果.Banknotes automatic sorting system can effectively improve work efficiency of a bank ,for auto‐matic banknote processing system ,recognition accuracy of banknotes automatic sorting system based on image processing technology is critical .Based on theoretical knowledge of machine vision and pattern rec‐ognition ,and combined with the features of RMB ,paper currency recognition based on image processing method ,based on the“rewards”feature extraction algorithm of linear sensor mechanism and serial number module method are proposed .After tilt correction ,local feature extraction ,gray enhancement and image block of banknotes image ,the above method is applied to an image processing system based on CIS sensor , experimental results show that the method has a high recognition accuracy ,achieves the desired effect of automatic banknotes sorting .

关 键 词：图像处理 纸币清分 序列号识别 线性感知算法 CIS传感器 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]