[1]刘华春.卷积神经网络在车牌识别中的应用研究[J].计算机技术与发展,2019,29(04):128-132.[doi:10. 3969 / j. issn. 1673-629X. 2019. 04. 026]
　LIU Hua-chun.Research on Application of Convolutional Neural Network in License Plate Recognition[J].,2019,29(04):128-132.[doi:10. 3969 / j. issn. 1673-629X. 2019. 04. 026]
点击复制

卷积神经网络在车牌识别中的应用研究()
分享到：

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

卷:
29
期数:
2019年04期
页码:
128-132
栏目:
应用开发研究
出版日期:
2019-04-10

文章信息/Info

Title:
Research on Application of Convolutional Neural Network in License Plate Recognition
文章编号:
1673-629X(2019)04-0128-05
作者:
刘华春
成都理工大学 工程技术学院 电子信息与计算机工程系,四川 乐山 614007
Author(s):
LIU Hua-chun
Department of Electronic Information and Computer Engineering,Engineering & Technical College of Chengdu University of Technology,Leshan 614007,China
关键词:
车牌识别; 卷积神经网络; 支持向量机; 改进LeNet-5卷积网络; 深度学习
Keywords:
license plate character recognition; convolutional neural network; support vector machines; improved LeNet-5 convolutional network; deep learning
分类号:
TP391
DOI:
10. 3969 / j. issn. 1673-629X. 2019. 04. 026
摘要:
为了改善传统车牌识别方法中过分依赖车牌特征和鲁棒性不强等问题,将具有良好图像识别性能的卷积神经网络LeNet-5模型引入到车牌字符识别中,并对其结构进行改进以满足需要。设计了2个网络分别进行汉字和数字/字母识别,将输出层类别由10增加到31和34;C5卷积层的特征面数目增加到480,输入图像像素增加到64×64。对改进后的网络进行了实验,并分别与3层BP神经网络和支持向量机（SVM）进行对比测试。实验结果表明,该卷积神经网络避免了传统车牌字符识别方法中复杂的特征提取,增强了鲁棒性,提高了准确率。改进后的LeNet-5相比BP神经网络在识别准确率上可提高约6%,识别速度也更快;与SVM相比较,汉字分类准确率可以提高约7%,字符/数字准确率可以提高约4%。
Abstract:
In order to improve the excessive dependence on license plate features and lack of robustness in traditional vehicle license platerecognition methods,the convolutional neural network LeNet-5 model with great image recognition performance is introduced into the license plate character recognition and its structure is improved to meet the needs. The main improvements are the design of two networks for Chinese character and digit/letter recognition,and increasing the output layer class from 10 to 31 and 34,and the number of featurefaces in the C5 convolutional layer to 480,and the input image pixels to 64×64. The improved network is tested and compared with a3-layer BP neural network and a support vector machine (SVM). The experiment shows that the convolutional neural network avoids the complex feature extraction in the traditional license plate character recognition method,and enhances the robustness and accuracy.Compared with BP neural network,the improved LeNet-5 accuracy can be improved by about 6% and the recognition speed is faster. Compared with SVM,Chinese character classification accuracy can be improved by about 7% ,and character/ digit accuracy can be increased by about 4%.

相似文献/References:

[1]邹星.一种基于模板库的车牌字符识别算法[J].计算机技术与发展,2010,(04):128.
　ZOU Xing.A License Plate Character Recognition Arithmetic Based on Template Library[J].,2010,(04):128.
[2]杨述斌 张阳.复杂车辆图像中的车牌快速形态定位算法[J].计算机技术与发展,2008,(06):50.
　YANG Shu-bin,ZHANG Yang.Fast Morphological Locating Algorithm of Vehicle License Plate in Complex Vehicle Images[J].,2008,(04):50.
[3]宓浩 张燕平.基于特征投影和交叉覆盖神经网络的车牌识别[J].计算机技术与发展,2007,(10):76.
　MI Hao,ZHANG Yan-ping.License Plate Location Based on Projection Character and Alternate Covering Neural Network[J].,2007,(04):76.
[4]李波 曾致远 周建中 罗勤.车牌识别系统研究与实现[J].计算机技术与发展,2006,(06):10.
　LI Bo,ZENG Zhi-yuan,ZHOU Jian-zhong,et al.Study and Realization for License Plate Recognition System[J].,2006,(04):10.
[5]孟晓莉 赵安军 马光思.基于数学形态学的车牌定位研究与实现[J].计算机技术与发展,2010,(11):84.
　MENG Xiao-li,ZHAO An-jun,MA Guang-si.Car License Plate Location Research and Implementation Based on Mathematical Morphology[J].,2010,(04):84.
[6]孔超 高珏 陈开云[] 许华虎[].基于多特征的车辆一致性检测的研究与应用[J].计算机技术与发展,2012,(03):31.
　KONG Chao,GAO Jue,CHEN Kai-yun,et al.Target Multi-Feature Based Consistency Checking of Vehicles[J].,2012,(04):31.
[7]霍玲玲,徐文军,赵云青,等.嵌入式车牌识别系统的设计与实现[J].计算机技术与发展,2013,(04):206.
　HUO Ling-ling,XU Wen-jun,ZHAO Yun-qing,et al.Design and Implementation of an Embedded System of Vehicle License Plate Recognition[J].,2013,(04):206.
[8]刘静.基于最小二乘支持向量机车牌字符特征识别[J].计算机技术与发展,2013,(05):195.
　LIU Jing.LSSVM-based License Plate Character Feature Recognition[J].,2013,(04):195.
[9]崔凤焦.表情识别算法研究进展与性能比较[J].计算机技术与发展,2018,28(02):145.[doi:10．3969/j．issn．1673－629X．2018．02．031]
　CUI Feng-jiao.Ｒesearch and Performance Comparison of Facial Expression Ｒecognition Algorithm[J].,2018,28(04):145.[doi:10．3969/j．issn．1673－629X．2018．02．031]
[10]张丹丹,李雷. 基于PCANet-RF的人脸检测系统[J].计算机技术与发展,2016,26(02):31.
　ZHANG Dan-dan,LI Lei. Face Detection System Based on PCANet-RF[J].,2016,26(04):31.
[11]刘华春,侯向宁.采用视觉显著性和卷积网络的车牌定位算法[J].计算机技术与发展,2019,29(06):74.[doi:10. 3969 / j. issn. 1673-629X. 2019. 06. 016]
　LIU Hua-chun,HOU Xiang-ning.A License Plate Location Algorithm Using Visual Saliency and Convolutional Network[J].,2019,29(04):74.[doi:10. 3969 / j. issn. 1673-629X. 2019. 06. 016]

更新日期/Last Update: 2019-04-10