Abstract: Traffic sign detection can be used in intelligent transportation system to improve the safety.Traffic signs have special shapes and colors，but most detecting methods ignores color information or just segmentated with const threshold values，which resulted in loss of adaptability and robustness.This paper presents an SVM based method using color and shape information for detection of traffic signs．This method is mainly carried are in two steps．First，it segments the image using SVM，then detects the traffic sign by SVM and the shape information．The methods by detecting blue designation traffic sign are tested.The experiments show the good robustness and high detection accuracy.

Key words: Traffic Sign Recognition（TSR）, image detection, Support Vector Machine（SVM）

摘要： 交通标志检测在智能交通系统中的作用是帮助驾驶提高安全性。交通标志都具有特定的颜色和形状，但是现有的检测方法大多使用固定阈值分割等非智能方法，缺乏自适应性和鲁棒性。使用支持向量机分割彩色交通标志图像，再结合形状特征，实现了一种新的智能检测方法；并以蓝色交通指示标志为检测对象，使用所提出的方法进行实验。实验结果表明，该方法鲁棒性好、检测准确率高。

关键词: 交通标志识别, 图像检测, 支持向量机