1 3
DOI 10.1007/s00530-014-0440-7
Multimedia Systems
SPECIAL ISSUE PAPER
Vehicle collision risk estimation based on RGB‑D camera
for urban road
Zhenyu Shan · Qianqian Zhu · Danna Zhao
© Springer-Verlag Berlin Heidelberg 2014
1 Introduction
In past decades, most cities have achieved new progress
in increment of vehicle number, which increases the inci-
dence of traffic accidents. In fact, most traffic accidents
can be avoided when drivers concentrate on the traffic
state and obey traffic regulations [1]. In view of this situ-
ation, what many traffic administrative departments do is
focusing on punishing illegal driving behaviors. It needs
lots of polices monitoring around the clock. To improve
the efficiency, varieties of sensors are installed on/under
the road or by the roadside to automatically detect the
various types of violations, such as red light running,
speeding driving and overtaking [2]. Besides, ignoring
safe distance is also an important factor in incurring col-
lision which often causes fatal losses. When at the same
speed, the closer the distance is, the higher the collision
risk is the safe distance in vehicle driving must be kept
to avoid collision. However, it has not obtained enough
attention by traffic administrative departments because it
is difficult to automatically estimate whether safe distance
is maintained [3].
Car radar installed in the vehicle is the most common
sensors for safe distance guarantee [4]. Other than one at
rear, more radars (usually four) are installed at the side of
vehicles in some kinds of high-end cars. It can be used to
remind the driver to keep a safe distance. However, there
are two limitations to punish the behavior of ignoring safe
distance by using it [5]. First, although the warning infor-
mation from the radar reveals the risk of collision, traffic
administrative departments have no access to it. Second,
the accuracy may not be guaranteed while the vehicle is
running because vehicle-borne radar is mostly designed
for vehicle reversing. Thus, the sensor or sensor network
installed on the road with convenient communications
Abstract Traffic violation is the main cause of traffic
accidents. To reduce the incidence of traffic accidents, the
common practice at present is to strength the penalties for
traffic violation. However, little attention has been paid to
issue warning for dangerous driving behaviors, especially
for the case where two vehicles have a good chance of col-
lision. In this paper, a framework for collision risk estima-
tion using RGB-D camera is proposed for vehicles running
on the urban road, where the depth information is fused
with the video information for accurate calculation of the
position and speed of the vehicles, two essential parame-
ters for motion trajectory estimation. Considering that the
motion trajectory or its differences can be considered as a
steady signal, a method based on autoregressive integrated
moving average (ARIMA) models is presented to predict
vehicle trajectory. Then, the collision risk is estimated
based on the predicted trajectory. The experiments are car-
ried out on the data from the real vehicles. The result shows
that the accuracy of position and speed estimation can be
guaranteed within urban road and the error of trajectory
prediction is very minor which is unlikely to have a sig-
nificant impact on calculating the probability of collision in
most situations, so the proposed framework is effective in
collision risk estimation.
Keywords Dangerous driving behavior · Collision risk ·
RGB-D camera · ARIMA
Z. Shan (*)
Intelligent Transportation and Information Security Lab,
Hangzhou Normal University, Hangzhou, Zhejiang, China
e-mail: shanzhenyu@zju.edu.cn
Z. Shan · Q. Zhu · D. Zhao
Hangzhou Normal University, Hangzhou, China