The Determination of Vehicular Trajectories Based on Video Image Processing Technology
Shuo Li, Xiu-jiang Zhao, Yan Gao, Jianlong Zhang · International Conference on Electric Technology and Civil Engineering · 2012
This paper represents a brand new method on studying the information of vehicles in running state on road. First of all, by modeling the road background by the model of updating the backgrounds of the video images of running vehicles on road continuously the initial video image data of vehicles in running state are obtained. Secondly, a new method which combines the simplified profile morphological operations with the band filter is then suggested to eliminate the vehicular shadows on the video images. And thirdly, a vehicular position tracking algorithm is proposed to identify the positions of different vehicles in running state via the resulted foreground video image data. After that, the continuous speed and acceleration rate information of running vehicles is obtained by applying the weighted window regression approach. All the trajectory curves on the time--space coordinates for all vehicles running on all lanes of the road of concern are therefore obtained.