基于颜色和形状的交通标志检测与分类  

Road Traffic Sign Detection and Classification Based On Color and Shape Features

作　　者：张瑞[1] 张朋[1] 

机构地区：[1]首都师范大学信息工程学院,北京100048

出　　处：《微计算机信息》2010年第35期226-228,共3页Control & Automation

摘　　要：本文旨在分析对比现有颜色分割算法的基础上,提出一种基于HSI颜色模型的颜色分割算法,该算法对在不同复杂背景下拍摄的图片,能够分割出人们感兴趣的颜色并排除大部分干扰颜色。在交通标志识别系统应用中,主要提取红、黄、蓝三种颜色,并且算法应能有效地排除其他颜色,尤其是绿色的干扰,可以接受一定的噪声影响。在此基础上,利用形状特征进行交通标志的检测与分类。实验结果表明,系统能实现复杂场景中交通标志的准确定位,但交通标志分类方法仍有待改进。Based on HSI color model,this paper seeks to propose a method for color segmentation on the basis of analyzing and comparing the existing color segmentation algorithms.Using the method,people can extract interested in the color and the exclusion of the majority of interference colors from the photographs with different complex background.Considering the traffic sign recognition system application,red,yellow and blue three colors are mainly extracted from image,and algorithm should be able to eliminate other colors effectively,especially green interference,can accept a certain impact of the noise.On the basis,the use of the shape features,traffic signs can be detected and classified.Experiments based on Matlab show that the system can detect traffic signs accurately under complex backgrounds,but the traffic signs classification method still needs to be improved.

关 键 词：交通标志检测与分类 颜色分割 拐角提取 神经网络 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]