基于SVM的字符验证码识别研究

Research on Recognition of Character Verification Code based on SVM

验证码广泛应用在互联网各个领域,也称为互联网安全防火墙,本质上验证码是为了区分人与机器人,防止网络机器人的攻击.然而随着OCR技术的发展,验证码暴露出来的安全问题也非常严峻.以高校系统中的字符型验证码作为研究对象,对验证码进行二值化,去噪,分割等操作,采用SVM支持向量机的方式对数字验证码进行识别.实验结果表明,当训练集到达一定数量时,对字符验证码具有较好的识别率.

Verification codes were widely used in various fields of the Internet,also known as Internet security firewalls.verification codes are used to distinguish between humans and robots to prevent attacks by network robots.However,with the development of OCR technology,the security problem exposed by the verification code is also very serious.In this paper,the character verification code in college system is taken as the research object,and the verification code is binarized,denoised,segmented,etc,and recognizing it based on SVM support vector machine.The experimental results show that when the training set reaches a certain number,it has a good recognition rate for the character verification code.