基于字符型验证码破解算法研究  

Research on Character Based Verification Code Decoding Algorithm

作　　者：张晓菲[1] 邹婷[1] 刘培鹤[1] 何文才[1] 

机构地区：[1]北京电子科技学院通信工程系,北京100070

出　　处：《北京电子科技学院学报》2016年第4期50-54,77,共6页Journal of Beijing Electronic Science And Technology Institute

摘　　要：验证码主要用于人机区分,在人机交互领域具有广泛的应用前景。对验证码破解算法的研究,可以实现验证码自动识别显示,使网络信息录入更加方便快捷;同时也可以防止验证码被暴力破解,减少验证码的生成漏洞。本文简要介绍了基于人工神经网络的字符识别方法,分析了验证码破解过程,对基于数字图像处理的相关算法和基本的验证码分割算法进行了分析,并提出了基于上下连通域字符分割法,提高了验证码破解的准确率和效率。The verification code is mainly used to distinguish human and computer, which has a wild range of application in the field of human computer interaction. The research on verification code decoding algorithm can realize the recognition of the verification code automatically to input the network information more convenient. In the meantime, it can prevent the verification code being brute with violence and reduce the vulnerability of verification code. This paper briefly introduces the method of character for identified that based on artificial neural network, analyzes the process of the code which is cracked and the related algorithms of digital image processing and verification code segmentation algorithm. W the crack efficiency and accuracy of ing verification code for connective regions between the points of up-contour and down-contour.

关 键 词：神经网络 图像处理 字符分割 验证码破解 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]