[1]李建平,王 钊.基于 PSO-CNN 的验证码识别算法研究[J].计算机技术与发展,2022,32(09):51-55.[doi:10. 3969 / j. issn. 1673-629X. 2022. 09. 008]
　LI Jian-ping,WANG Zhao.Research on Verification Code Recognition Algorithm Based on PSO-CNN[J].,2022,32(09):51-55.[doi:10. 3969 / j. issn. 1673-629X. 2022. 09. 008]
点击复制

基于 PSO-CNN 的验证码识别算法研究()

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:
32
期数:
2022年09期
页码:
51-55
栏目:
媒体计算
出版日期:
2022-09-10

文章信息/Info

Title:
Research on Verification Code Recognition Algorithm Based on PSO-CNN
文章编号:
1673-629X(2022)09-0051-05
作者:
李建平; 王 钊
东北石油大学 计算机与信息技术学院,黑龙江 大庆 163318
Author(s):
LI Jian-ping; WANG Zhao
School of Computer and Information Technology,Northeast Petroleum University,Daqing 163318,China
关键词:
粒子群优化算法; 卷积神经网络; 验证码; 数据预处理; Tesseract
Keywords:
particle swarm optimization ( PSO) ; convolutional neural network ( CNN) ; verification code; data preprocessing; Tesseract
分类号:
TP391
DOI:
10. 3969 / j. issn. 1673-629X. 2022. 09. 008
摘要:
伴随着互联网的高速发展,非法用户恶意攻击网站、恶意注册、暴力破解用户密码等事件也随之而来。 为了解决这些网络安全问题,作为网络安全第一道防线的验证码技术应运而生。 但在实现自动登录合同管理系统的过程中,验证码自动化识别一直是个技术难点,验证码自动化识别准确率直接影响了业务处理效率,故此提出了一种基于 PSO-CNN 的验证码识别方案。 针对一万张验证码图片的数据集进行灰度化、二值化以及降噪三步数据预处理之后,通过 PSO 优化算法在卷积神经网络训练数据集的过程中找出最佳的网络层数和卷积核大小。 经过反复的实验,结果表明基于 PSO-CNN的验证码识别算法对数字与字母混合验证码识别准确率可达 96. 26% ,为合同管理系统实现自动登录提供了可靠的技术支持。
Abstract:
With the rapid development of Internet,illegal users malicious attacks on websites,malicious registration,violent cracking of user passwords? and other events have followed. In order to solve these network security problems,verification code technology,as thefirst line of defense? ? ? ? of network security, came into being. However, in the process of realizing automatic login to the contract management system, automatic identification of verification code has always been a technical difficulty. The automatic identification accuracy ofverification code directly affects the business processing efficiency. There fore,a verification code recognition scheme based on PSO -CNN is proposed. After the three - step data preprocessing of grayscale, binarization and noise reduction for 10000 verification codeimages,the best number of network layers and convolution kernel size are found in the process of convolutional neural network training data set through PSO optimization algorithm. The results of repeated experiments show that the recognition accuracy of mixed digital and letter verification code based on PSO-CNN has reached 96. 26% ,which provides reliable technical support for automatic login of system.

相似文献/References:

[1]张雯雰 李丽娟 滕少华[] 罗玉玲.粒子群优化算法在桁架结构优化中的应用[J].计算机技术与发展,2010,(05):223.
　ZHANG Wen-fen,LI Li-juan,TENG Shao-hua,et al.Improved Particle Swarm Optimizer Algorithm for Design Optimization of Structures[J].,2010,(09):223.
[2]张家柏 王小玲.基于聚类和二进制PSO的特征选择[J].计算机技术与发展,2010,(06):25.
　ZHANG Jia-bai,WANG Xiao-ling.A Novel Algorithm Based on K-Means Clustering and Binary Particle Swarm Optimization[J].,2010,(09):25.
[3]张艳丽 保文星.粒子群优化算法在图像边缘检测中的研究应用[J].计算机技术与发展,2009,(05):26.
　ZHANG Yan-li,BAO Wen-xing.Research and Application of Image Edge Detection Based on PSO Algorithm[J].,2009,(09):26.
[4]聂笃宪.基于PSO自适应正则化参数图像恢复的研究[J].计算机技术与发展,2009,(01):106.
　NIE Du-xian.Research on Adaptively Regularized Parameter Image Restoration PSO- Based[J].,2009,(09):106.
[5]谭伟 李向.微粒群优化算法的研究[J].计算机技术与发展,2009,(03):87.
　TAN Wei,LI Xiang.Research Status and Development of Particle Swarm Optimization[J].,2009,(09):87.
[6]曾万里 危韧勇 陈红玲.基于改进PSO算法的BP神经网络的应用研究[J].计算机技术与发展,2008,(04):49.
　ZENG Wan-li,WEI Ren-yong,CHEN Hong-ling.Research and Application of BP Neural Network Based on Improved PSO Algorithm[J].,2008,(09):49.
[7]朱玉平.一种改进粒子群优化算法[J].计算机技术与发展,2008,(11):106.
　ZHU Yu-ping.An Algorithm of Modified Particle Swarm Optimization[J].,2008,(09):106.
[8]林杰 孙淑霞 文武.基于粒子群优化算法的图像小波阈值去噪研究[J].计算机技术与发展,2007,(04):204.
　LIN Jie,SUN Shu-xia,WEN Wu.An Image Denoising Method Based on Wavelet Transform and Particle Swarm Optimization[J].,2007,(09):204.
[9]肖裕权 周肆清.基于粒子群优化算法的数据流聚类算法[J].计算机技术与发展,2011,(10):43.
　XIAO Yu-quan,ZHOU Si-qing.Clustering Evolving Data Streams Based on Particle Swarm Optimization[J].,2011,(09):43.
[10]林云光 陈月辉 邵光亭.基于前馈人工神经网络的miRNA预测[J].计算机技术与发展,2012,(05):19.
　LIN Yun-guang,CHEN Yue-hui,SHAO Guang-ting.Prediction of miRNA Based on Feedforward Artificial Neural Network[J].,2012,(09):19.

更新日期/Last Update: 2022-09-10