[1]王洪元,刘忠杰,刘爱萍.基于多特征组合的车牌识别算法[J].常州大学学报(自然科学版),2010,(04):47-51.
　WANG Hong- yuan,LIU Zhong- jie,LIU Ai- ping.Research of License Plate Recognition Algorithm Based on Multi- Feature Combination[J].Journal of Changzhou University(Natural Science Edition),2010,(04):47-51.
点击复制

基于多特征组合的车牌识别算法()
分享到：

常州大学学报(自然科学版)[ISSN:2095-0411/CN:32-1822/N]

卷:
期数:
2010年04期
页码:
47-51
栏目:
出版日期:
2010-12-25

文章信息/Info

Title:
Research of License Plate Recognition Algorithm Based on Multi- Feature Combination
作者:
王洪元;  刘忠杰;  刘爱萍
常州大学信息科学与工程学院, 江苏 常州 213164
Author(s):
WANG Hong- yuan;  LIU Zhong- jie;  LIU Ai- ping
School o f Informat ion Science and Eng ineering, Changzhou U niv ersity, Chang zhou 213164, China
关键词:
图像预处理;  车牌定位;  车牌字符分割;  车牌字符识别
Keywords:
image preprocessing;  license plate location;  license plate character segmentat io n;  license plate character reco gnition
分类号:
T P 391^43
文献标志码:
A
摘要:
车牌识别是智能交通系统中的重要组成部分, 主要包括车牌定位、车牌字符分割和车牌字符识别3 部分。本文使用 M ATLAB 对车牌识别进行了研究, 提出了一种基于多特征组合的车牌识别算法。该算法首先通过边缘检测算子和数学形态学 运算得到车牌候选区, 然后利用车牌的矩形特征、宽高比特征和垂直投影特征对这些候选区域进行筛选, 并定位出车牌; 二值 化、滤波和形态学开运算后使用投影二分法分割出7 个车牌字符; 最后使用模板匹配和特征统计相结合的方法识别出车牌字 符。实验表明该算法是有效的、可行的, 能够在复杂背景下准确的定位和识别出车牌字符。
Abstract:
License plate recog nit ion play s an impor tant role in Intelligent T ranspo rtat ion System, mainly including license plate locat ion, license plate character segmentat ion and license plate character recogn-i t ion. This paper studied the license plate recog nit ion w ith MA TLAB, and pro posed a license plate recogn-i t ion alg orithm based o n multi- featur es. T his metho d f irst ly conf irmed v ehicle license candidate r eg io ns w ith edge detect ion o perator and mathemat ical mo rpholo gy o perat io ns, then screened these candidate reg ions w ith license plate r ectangular features, aspect rat io features and char acterist ics o f the vertical pro ject ion, and located license plate. After binarizat ion, filtering and mo rpholo gical o pening operat ion, seven l-i cense plate characters w ere segmented by project ion dichotomy. Finally, license plate character w as recognized by the combining method of template matching and statistical characterist ics. Experiment demonst rates that this method is feasible, ef fect ive and can accurately locate and identify the license plate characters in the complex backgro und.

参考文献/References:

[1] Zhan g C heng, S un Guan g- m in, Chen Deming. A rapid l ocat ing m ethod of veh icle licens e pl at e based on charact erist ics of charact er sc conn ect ion and projection [C] MProc of th e S econ d IEEE Conf erence on Indus t rial Elect ronics and Appli cat ions . New York: Sprin ger- Verlag, 2007: 2 546- 2 549.
[2] Bai H ong- liang, Liu Ch ang - ping. A hybrid licens e plat e ext raction method based on edge s tat ist ics and m orphology [C] M Proc of th e 17th Int ern at ional on Pat t ern Recogn iti on . Wash ingt on : John- Part ridge, 2004: 831- 8341
[3] Guo J ing- Min g, Liu Yun- Fu. License plate localizat ion an d charact er s egment at ion w ith f eedback self- learnin g and hyb rid binarizat ion t echniques [J] . T ran sacti on s on Vehicular Technology, 2008, 57 ( 3) : 1 417- 1 424.
[4] Chang Shyang - Lih , Chen Li- S hien , Chun g Ytm - Chun g. Aut omati c l icen se plat e recognit ion [J] . IEE E T ransact ions on Int ell igent T rans portat ion Syst em , 2004, 5 ( 1) : 42- 53.
[5] 冼允廷, 路小波, 施毅, 等1 基于投影二分法的车牌字符分割 方法[J] 1 交通计算机, 2007, 25 ( 5) : 69- 71.
[6] 李孟歆, 吴成东1 基于分级网络的车牌字符识别算法[J] 1 计算机应用研究, 2009, 26 ( 5) : 1 703- 1 7041

备注/Memo

备注/Memo:
基金项目: 江苏省高校自然科学基金( 08KJB520002) 作者简介: 王洪元( 1962- ) , 男, 江苏长熟人, 教授, 主要研究方向图像处理、模式识别、人工智能。
更新日期/Last Update: 2010-12-25