认证码字符识别方法的研究  

An Approach to Recognition of Authentication Code

作　　者：张淑雅[1] 赵一鸣[1] 赵晓宇[1] 李均利[1] 

机构地区：[1]宁波大学数字技术与应用软件研究所,浙江宁波315211

出　　处：《宁波大学学报（理工版）》2007年第4期429-433,共5页Journal of Ningbo University：Natural Science and Engineering Edition

基　　金：国家自然科学基金(60302012);浙江省自然科学基金(Y106505)

摘　　要：从加强网站安全性的角度,研究认证码字符识别的技术要点.首先,提出预处理算法去除图像中的干扰因素,分割出字符;其次,采用分类器技术对预处理后得到的字符进行识别.实验指出有效的预处理算法和选择合适分类器对于快速准确识别认证码起着至关重要的作用,证实了目前一些认证码在保障网站安全性的同时,尚存在一些漏洞,需要继续完善。Recognition of authentication code is investigated from the perspective of strengthening website security. A preprocessing algorithm is first developed to remove the disturbance factors in the image and obtain the characters, followed by using some classifiers for recognition purposes. The experimental results suggest that the developed algorithm and selected classifiers are the most critical part in the recognition process. Some loopholes are still identified in the authentication codes in the website protection phase, and hence further investigation into this issue is needed.

关 键 词：认证码 字符识别 图像预处理 分类器 

分 类 号：TP391.43[自动化与计算机技术—计算机应用技术]