自然场景下三角形交通标志的检测与识别

根据三角形交通标志的颜色和形状特征，提出了一种适用于自然场景下三角形交通标志的检测与识别方法. 该方法首先利用颜色分割粗略提取标志区；其次提取标志区轮廓边缘和直线拟合，确定三角形标志的3个顶点，精确检测出完整的三角形标志区；最后设计一种分块特征提取方法对检测出的三角形标志和所有参考三角形标志进行特征提取，通过特征匹配识别出三角形交通标志类别. 实验结果表明，所提出的检测与识别方法能更有效地识别自然环境中三角形交通标志方法，且适用性强.

Detection and Recognition of Triangular Traffic Signs in Natural Scenes

An approach is proposed to detect and recognize triangular traffic signs in natural scenes according to color and geometric features of the signs. The sign is first roughly extracted based on color segmentation.A straight line fitting method is used to detect the three sides of the triangular signs to obtain the complete triangular sign. A partitioning feature method is used to obtain feature vectors from all reference triangular traffic signs as well as the detected ones, then the detected triangular traffic sign is recognized by feature vector matching. Experimental results show that the proposed method is effective for recognizing triangular traffic signs in natural scenes.