基于ICA降维的车牌汉字识别研究  

Research of license plate Chinese characters recognition based on ICA

作　　者：沈洋[1,2] 詹永照[1] 单士娟[2] 

机构地区：[1]江苏大学计算机科学与通信工程学院,江苏镇江212013 [2]宿迁学院计算机科学系,江苏宿迁223800

出　　处：《计算机工程与设计》2012年第3期1127-1131,共5页Computer Engineering and Design

基　　金：江苏省自然科学基金项目(BK2009199)

摘　　要：针对目前复杂环境下车牌汉字图像识别率较低,识别时间较长等问题,提出了一种基于伪Zernike矩和独立主成分分析(ICA)的改进概率神经网络(PNN)车牌汉字识别方法。该方法是将车牌汉字图像的伪Zernike矩通过独立主成分分析降维,再将降维后的特征输入所提出的一种基于代表点的改进概率神经网络中进行训练和识别,从而有效地实现车牌汉字的识别。将该方法应用于复杂环境下的车牌汉字图像识别实验,实验结果表明,该方法能有效地降低特征维数,减少识别时间,并能显著地提高车牌汉字的识别率。Aimed at the low recognition rate and speed of Chinese characters in complex environment, an improved probabilistic neural network （PNN） recognition algorithm of license plate Chinese characters based on pseudo-Zernike moment and independent component analysis （ICA） is presented. In this algorithm, the pseudo-Zernike moment＇s dimension of Chinese character image are reduced by ICA, then the obtained features are regarded as the input of an improved PNN, which is adopted to recognize Chinese characters. Thereby Chinese characters recognition is effectively implemented. This algorithm is applied to Chinese characters recognition in complex environment. The experimental results show that the algorithm can effectively reduce the dimension of feature, decrease the recognition time and raise the accuracy of recognition.

关 键 词：伪ZERNIKE矩 独立主成分分析 图像特征降维 概率神经网络 车牌汉字识别 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]