To meet the demands of real-time intelligent transport system and to make image-pretreatment promptly, this paper introduced a new method for extracting traffic sign, example of traffic ban sign. First, the method segmented a color image to binary image with RGB color information. Then we detected the edge of the object. The circumference and the area of the thinning-image was gotten by chain code. And then we can get the circularity of the target. At last the method used the optimal threshold in circularity to extract the traffic ban sign. We have 20 sample images to test the method. It's cost less than 0.3 second for each sample extraction. Experiments show that this method has less computation and more efficiency.%为满足智能交通系统的实时性要求,提高交通标志图像预处理的速度,以交通禁令标志为例,介绍了一种提取交通标志的新方法.该方法首先通过RGB彩色信息快速分割彩色标志图像转换为灰度图像,使用log边缘提取算子进行图形的边缘检测.对细化后的灰度图像,再利用链码计算图形周长和面积,进而求取目标圆形度,设置合适的面积和圆形度阈值可提取出交通标志.使用20个自拍样本测试,捉取用时均小于0.3s.实验结果表明本方法计算简单,实时性好,检测的有效性较高.
展开▼
