多车道路面的大车车型分类特征及提取方法

针对车流量大的多车道路面，提出了以车体结构和颜色为基础的6个能有效区分大客车和大货车的特征，并介绍了提取过程。该过程首先使用Sobel算子与颜色融合的抗干扰水平边缘检测方法提取区间分界线，然后以基于RGB和HSI相结合的颜色分类方法识别区间颜色，进而获得特征的表示。现场实测结果表明，这些特征提取准确率高且速度快，能满足实时性要求。基于这些特征的车型分类系统已应用于实际现场并取得了良好的识别效果。

Big vehicle classification features and their extraction method for multi-lanes roads

Concerning the multi-lanes road with heavy traffic flow,six features related to vehicle's structural region and color were proposed,which can distinguish truck from bus efficiently,and the features extracting process was introduced.First the boundary of regions was detected by using an insensitive noise edge detection method based on the fusion of Sobel operator and color,then the region color was recognized according to a color classification approach combining RGB and HSI color spaces,and the delineation ...