机构地区: 重庆邮电大学计算机科学与技术学院计算机科学与技术研究所
出 处: 《小型微型计算机系统》 2012年第12期2709-2714,共6页
摘 要: 限速标志识别系统是智能交通系统的一个重要组成部分,它能有效地辅助司机安全驾驶.针对限速标志的数字字符识别问题,提出一种基于超网络模型的模式识别方法.首先介绍了超网络计算模型及其分类原理;然后采用颜色分割和形状分析相结合的方法对限速标志进行定位,并提取出限速数字字符特征;最后以限速字符的特征向量为训练样本对超网络模型进行演化学习.本文使用超网络模型对限速标志20、40、60、80 km/h进行识别.实验结果表明,基于超网络模型的道路限速标志识别系统最快只需3次迭代便可以完成对样本的学习,识别率为96.15%.和其它传统模式识别方法相比,该模型具有学习时间短、识别率高的优点,为解决现实应用中的道路限速标志识别问题提供了可能. Road speed limit sign recognition system is an import part of intelligent transportation system, which effectively improves safety driving. To solve the characters recognition problem of the speed limit signs, a hypemetwork model is proposed in this paper. Firstly, we introduce the hypernetwork model and the realization of a hypernetwork classifier in detail. And then we employ the color segmentation and shape analysis techniques to locate the speed limit signs and extract their features. The extracted features are used as training set for the hypernetworks learning. Finally, hypernetworks is used to recognize the speed signs: 20, 40, 60 and 80 km/h. Experimental results show that the recognition rate of the proposed hypernetworks-based road speed limit sign recognition system is 96.15%. The system learning process takes only 3 iterations. Compared with other traditional recognition methods, hypernetworks is time-saving and provides a high recognition performance, which makes hypernetwork very suitable for solving speed limit signs recog- nition oroblems in real world.
关 键 词: 限速标志识别 智能交通系统 超图 超边 超网络
领 域: [自动化与计算机技术] [自动化与计算机技术]