一种前方车辆后轮接地点检测算法

A Detection Algorithm of Rear Wheel Ground Touching Point of Vehicle in Front

现有车辆定位算法通常通过检测车辆底部阴影来定位车辆,该算法受阴影形状大小变化的影响较大,定位精度较低。该文提出了一种通过检测前方车辆后轮接地点来精确定位车辆的方法。首先通过垂直边缘和对称性测度实现路面车辆的初定位;然后利用车辆垂直边缘和后部水平边缘显著性的特点获得后车轮可能存在的感兴趣区(ROI);最后在该ROI中通过Tamura纹理特征的粗糙度对车轮和地面进行区分,精确地得到车轮接地点位置。

The existed algorithm for on road vehicle location is often based on the shadow under vehicles which often be affected by shape and size of the shadow which will cause inaccuracy of vehicle location.To solve this problem,this paper proposes a new way which use rear wheel ground touching point to accurately locate the vehicle.This algorithm firstly uses edge information and symmetry to preliminarily get the position of the vehicle,then the vertical edge and horizontal edge are used to generate region of interesting（ROI） of rear wheels,finally,coarseness in Tamura texture features is used to separate wheel and ground so that the rear wheel ground touching point can be detected.