作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2012, Vol. 38 ›› Issue (13): 192-195,198. doi: 10.3969/j.issn.1000-3428.2012.13.057

• 人工智能及识别技术 • 上一篇    下一篇

基于正交盖氏矩和SVM的车牌字符识别

王桂文,孙 涵   

  1. (南京航空航天大学计算机科学与技术学院,南京 210016)
  • 收稿日期:2011-11-08 出版日期:2012-07-05 发布日期:2012-07-05
  • 作者简介:王桂文(1988-),女,硕士,主研方向:图像处理,计算机视觉,人工智能;孙 涵,副教授、博士

License Plate Character Recognition Based on Orthogonal Gegenbauer Moment and SVM

WANG Gui-wen, SUN Han   

  1. (College of Computer Science and Technology, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China)
  • Received:2011-11-08 Online:2012-07-05 Published:2012-07-05

摘要: 针对传统字符特征提取算法中特征不稳定的缺点,提出一种基于正交盖氏矩的特征提取方法。采用支持向量机解决车牌字符识别问题,自动寻找对分类有较好区分能力的支持向量,由此构成的分类器可以最大化类间间隔,达到正确区分类别的目的。实验结果表明,该方法对于实时视频流中的车牌识别能取得理想效果,在解决有限样本、非线性及高维模式识别问题中表现出优越的性能,且具有适应性强和效率高的特点。

关键词: 盖氏矩, 特征提取, 字符识别, 支持向量机, 分类器, 模式识别

Abstract: Aiming at the problem that the character features which are got by traditional feature extraction algorithm are not stable, this paper puts forward a feature extraction method based on orthogonal Gegenbauer moment. By using Support Vector Machine(SVM) method to solve the license plate character recognition problem, SVM can automatically search for classification which has good ability to distinguish between the support vector. The classifier can maximize kind of interval, and distinguish the purpose of the category. Experimental results show that this method can make the ideal effect in real-time streaming video of the license plate identification. In solving nonlinear finite sample, and high dimensional pattern recognition problem, it shows many special superior performance, and has strong adaptability and the characteristics of high efficiency.

Key words: Gegenbauer moment, feature extraction, character recognition, Support Vector Machine(SVM), classifier, pattern recognition

中图分类号: