Publication Details
Visual Analysis of Vehicle Trajectories for Determining Cross-Sectional Load Density
Špaňhel Jakub, Ing., Ph.D. (DCGM FIT BUT)
Sochor Jakub, Ing. (DCGM FIT BUT)
Herout Adam, prof. Ing., Ph.D. (DCGM FIT BUT)
Novák Jan, Ing., Ph.D. (TRC)
Road Safety, Lane Marking, Trajectory Analysis, Computer Vision, Vehicle Tracking
The goal of this work was to analyze the behavior of drivers on third class roads with and without horizontal lane marking. The roads have low traffic volume, and therefore a conventional short-term study would not be able to provide enough data. We used recording devices for long-term (weeks) recording of the traffic and designed a system for analyzing the trajectories of the vehicles by means of computer vision. We collected a dataset at 6 distinct locations, containing 1 010 hours of day-time video. In this dataset, we tracked over 12 000 cars and analyzed their trajectories. The results show that the selected approach is functional and provides information that would be difficult to mine otherwise. After application of the horizontal markings, the drivers slowed down and shifted slightly towards the outer side of the curve.
@ARTICLE{FITPUB12024,
author = "Roman Jur\'{a}nek and Jakub \v{S}pa\v{n}hel and Jakub Sochor and Adam Herout and Jan Nov\'{a}k",
title = "Visual Analysis of Vehicle Trajectories for Determining Cross-Sectional Load Density",
pages = "50--57",
booktitle = "Transactions on Transport Sciences",
journal = "Transactions on Transport Sciences",
volume = 10,
number = 1,
year = 2019,
publisher = "Palacky University",
ISSN = "1802-9876",
doi = "10.5507/tots.2019.002",
language = "english",
url = "https://www.fit.vut.cz/research/publication/12024"
}