基于边缘点运动估计的交通参数检测

Particle Swarm Optimization Based Image Matching of Moving Vehicles

利用运动车辆的边缘点包含了运动的位置及运动方向的特点,本文提出一种基于边缘点运动估计的视频检测方法,通过对前后两帧交通图像的运动边缘点在虚拟线圈内进行匹配,可以确定运动车辆在连续两帧图像中的运动矢量,从而获取车流、车速等交通参数。该方法大大减少了运算量,确保了视频检测的快速性和准确性。

As edge points of a moving vehicle contain the moving position and direction, this article proposes an arithmetic of video detection which is based on movement estimation of edge points. We can calculate the motion vector between two seriate images by matching the edge points in a virtual loop, thus gain the volume and speed of a vehicle. This method can reduce the operation quantity, and guarantee the rapidity and accuracy of video detection.