一种基于SURF的交通标志识别算法

A traffic sign recognition algorithm based on SURF

提出了一种基于SURF的交通标志识别算法。算法首先对从视频中抽取的图像进行预处理,之后使用交通标志的颜色与形状特征信息来检测与分割交通标志。然后使用SURF特征提取算法来抽取和描述已经定位的交通标志的特征值。最后,使用基于加权欧几里德距离的最近邻搜索方法在经过粗分类的特征模板库中进行搜索匹配。实验结果显示该算法具有较好的识别精度和速度。

In this paper, a traffic sign recognition algorithm based on SURF （Speeded Up Robust Features） is presented. Firstly, the algorithm used the color and shape features of traffic signs to detect and segment the traffic signs in the images after the image preprocessing. And then, eigen-values of the traffic signs which had been located from complicated background is extracted and descripted by SURF. Finally, on the recognition stage, a template match method besed on KNN-Seareh was used to get the best macther on the classfied eigen-database. The experiment results show high recognition accuracy and speed for the selected dataset.