光学技术, 2012, 38 (5): 564, 网络出版: 2012-10-08
多特征融合的交通标志检测与分类研究
The detection and classification of traffic signs by multi-feature fusion
图像识别 RGB模型 颜色分割 Canny边缘检测 形状识别 images recognition RGB model color segmentation Canny edge detection shape recognition
摘要
针对国内交通标志的属性特征, 提出了一种基于颜色和形状的交通标志分类算法。采用RGB模型分量值互减的方法对自然背景下的交通标志图像进行分割, 提取出感兴趣区域, 对标志区域进行形态学处理、边缘检测, 提取标志外层轮廓。利用Hough变换识别交通标志线性特征。在识别交通标志颜色和几何形状的基础上, 利用交通标志的分类知识实现交通标志的快速分类。实验结果表明, 该方法容易实现, 满足实时性要求, 并能达到较好的分类效果。
Abstract
For the special feature of China traffic signs, a new classification method based on color and shape for traffic signs is proposed. An improved RGB image segmentation algorithm is used to segment traffic signs from natural scenes, and the outermost contour of the signs region is extracted by using morphological processing and edge detection. The linear feature of traffic signs is recognized through Hough transform. The fast classification decision of traffic signs is made by integrating the information of color and geometric shape features of the traffic sign. Experimental results show that the algorithm is effective and can achieve better classification results.
李阳, 丁辉, 王云飞, 王怡蕾. 多特征融合的交通标志检测与分类研究[J]. 光学技术, 2012, 38(5): 564. LI Yang, DING Hui, WANG Yunfei, WANG Yilei. The detection and classification of traffic signs by multi-feature fusion[J]. Optical Technique, 2012, 38(5): 564.