基于序列图像的车载测速方法与实验

Vehicle-borne speed measurement method and experiment based on image sequences

为精确测量车辆的运动速度,文章提出了基于路面序列图像速度测量方法,该方法利用车载图像采集系统,在连续成像的条件下,实时采集运动路面的灰度图像,根据相邻帧路面图像的相似性,以灰度投影算法估计出帧间图像位移,同时结合图像采集系统的标定参数及帧频,计算车辆的运动速度。实验结果表明,该算法对纹理随机分布的平坦路面上运动的车辆速度的测量具有很好的稳定性,实验测得的运动速度与实际平均速度的误差在4%以内。

In order to accurately measure the movement speed of the vehicle,a speed measurement method based on the image sequences of the pavement is presented.This method uses the vehicle-borne image collection system to collect the gray image of the pavement in real time under the condition of continuous imaging.According to the similarity of the pavement images with adjacent frames,the interframe image displacement is estimated by using the gray projection algorithm,and the movement speed of vehicles is calculated based on the calibration parameters and frame rate of the image collection system.The experimental results show that this algorithm has good stability for the vehicle speed measurement on a flat road with randomly distributed texture,and the experimental measurement error between the motor instantaneous speed and the actual average speed is less than 4%.