信用卡数字识别研究

文字作为社会文明的载体,在信息科技领域占有重要的地位。光学字符识别(OCR)可以减轻人力繁琐的工作,本文重点研究实现了银行卡上印刷体字符识别的算法。首先进行银行卡字符识别的前期处理工作。采用常见的图像的预处理方法,如用高斯滤波器对银行卡图像进行了图像的复原去噪、二值化,以及倾斜校正等。然后对银行卡图像的字符区域进行了进一步的提取,结合银行卡图像的版面分析、字符区域投影操作,对印刷体字符进行了字符细化与分割。随后研究总结了字符的特征提取与识别算法。对分割后的印刷体字符进行了归一化处理,最终选用基于模板匹配的字符识别方法,并对银行卡图像进行特征提取后就以上方法进行了仿真实现,对识别结果进行了分析,其识别率较高,但仍有不理想的结果出现,需要改进。

As the carrier of social civilization,text occupies an important position in the field of information technology.Optical character recognition(OCR)can reduce the laborious work,this paper focuses on the realization of printed character recognition algorithm on bank CARDS.Firstly,the preliminary processing of character recognition of bank card is carried out.Common image preprocessing methods are used,such as gaussian filter for image restoration,denoising,binarization and tilt correction.Then the character area of the bankcard image is further extracted,and the printed characters are refined and segmented by combining the layout analysis of the bankcard image and the projection operation of the character area.Then the algorithm of character feature extraction and recognition is summarized.Finally,the character recognition method based on template matching is selected.After the feature extraction of bank card image,the above methods are simulated and the recognition results are analyzed.The recognition rate is relatively high,but there are still some unsatisfactory results that need to be improved.