提出了一种基于视频流的车辆速度估计方法.该方法首先使用光流分析得到车辆移动特征点,进而计算车辆运动矢量.然后通过对车辆运动矢量的统计得到运动矢量直方图,运动矢量直方图即可反映车辆运动速度的大小和方向.在建立车速估计函数的基础上,给出了算法的实现过程.实验结果表明该估计算法具有稳定性,是一种车速估计的有效方法.
This paper presents a vehicle speed estimation method on the basis of streaming video.First,the method analyzes optical flow to get the movement feature points of vehicles and then to calculate the vehicle motion vector.Followed,we obtain motion vector histogram of motion vector by a statistics means on them.Obviously,the motion vector histogram could represent the size and direction of the vehicle velocity.In this paper,we build a speed estimation function and give the implementation process of the algorithm,and the experimental results show that our algorithm has stability,and is effective on vehicle speed estimation.