摘要
针对银行卡数字卡号定位与识别的问题,本文提出一种基于模式匹配的银行卡数字卡号定位与识别算法。通过对银行卡数字图像进行图像灰度化、图像二值化、形态学操作、边缘检测、提取轮廓及区域定位等操作,对银行卡数字进行精准定位,对定位后的银行卡数字区域进行图像分割,将银行卡中分离的每一个数字图像分别与模板图像通过相关系数匹配法进行匹配,最终得到经过识别后对应的银行卡数字。在测试数据集上喷涂型银行卡数字识别正确率为94.4%,激凸印刷型银行卡数字识别正确率为89.5%,在总体样本中银行卡数字识别正确率为91.9%。经过测试可得出该算法能够有效地识别银行卡卡号,具有一定的市场价值。
Aiming at the problem of bank card digital card number location and recognition, this paper proposes a pattern matching-based algorithm for locating and identifying the digital card numbers of bank cards. Accurate positioning of bank card digits through image greyscale, image binarization, morphological operations, edge detection, extraction of contours and regional positioning of bank card digital images, and perform image segmentation on the bank card digital area after positioning.Through the correlation coefficient matching method,each digital image separated in the bank card is matched with the template image, and finally the corresponding bank card number after identification is obtained. On the test data set, the accuracy rate of the spray-type bank card number recognition was 94.4%, and the accuracy rate of the laser-printed bank card number recognition was 89.5%. In the overall sample, the accuracy rate of bank card number recognition was 91.9%. After testing, it can be concluded that the algorithm can effectively identify bank card numbers and has some market value.
作者
贾树林
郭磊
马双宝
董玉婕
林巍
JIA Shu-lin;GUO Lei;MA Shuang-bao;DONG Yu-jie;LIN Wei(School of Mechanical Engineering and Automation,Wuhan Textile University,Wuhan Hubei 430200,China;The Second Military Representative Office of the Military Representative Office of the Army Equipment Department in Wuhan,Wuhan Hubei 430200,China)
出处
《武汉纺织大学学报》
2021年第6期60-65,共6页
Journal of Wuhan Textile University
基金
大学生创新创业项目(202010495017,S202010495050)
武汉纺织大学研究生教学改革与研究项目(2021-2)
湖北省高校学生工作精品项目和实践育人特色项目(2019XGJPB2009)。
关键词
银行卡数字识别
模式匹配
图像分割
边缘检测
bank card digital recognition
image processing
image segmentation
edge detection
