Database Reference
In-Depth Information
Every year road accidents cause loss of lives, loss of money and serious delays on
city streets and freeways. Approximately 50 %
60 % of the delays on urban freeways
are caused by road accidents [
6
]. Reducing delays by faster accident analysis is a
challenging task. This task includes building an evidence collection system that can help
the police and city authorities to make more accurate and faster decisions. The best way
to achieve it is to provide video recordings from different angles and sources to the
decision maker. If such videos are available, cars that participated in an accident can be
faster moved out of the road and unblock traf
-
cient system for
collecting reports about road (and other) problems from citizens in an important part too.
During last decade vehicle-mounted surveillance cameras (VMSC) gained signif-
icant popularity [
7
]. VMSC can also be referenced as cars video registers or black
boxes or dash cams. They can help to prove innocence in case of a road accident or just
make a video of something interesting and uncommon. By now, VMSCs already have
a feature of annotating video with street name, as they have a GPS module onboard.
The most modern VMSCs have Internet access. It can be predicted, that in future most
of such smart cameras will have Internet access. VMSCs are produced by various
manufacturers, for example, Garmin,
1
Prestigio,
2
HP.
3
A typical VMSC is shown in
Fig.
1
(left). Millions of surveillance cameras are deployed, many of them are private.
Also, many people just make recordings with their smartphones. As a smartphone has a
camera, it can easily act as a VMSC if it is
c. A modern and ef
xed on a front window. Many smartphone
apps are available in Google Play and Apple AppStore. Examples of such applications
include: DailyRoads Voyager,
4
Axel Voyager,
5
AVR,
6
AutoBoy BlackBox,
7
CaroO
Pro
8
etc. their characteristics will be briefly discussed in the
'
related work
'
section.
A smartphone acting as a VMSC is shown in Fig.
1
(right).
Eventually, the cloud storage would keep videos of many events that can be of
interest for city management, police, road services and other governmental organiza-
tions. The challenge is in retrieving relevant video streams when they are really needed.
Owners of the surveillance camera may not be informed, that an accident happened just
in his/her camera view. Owner of a VMSC could pass by without stopping, when an
accident happens in front of their car. In such a situation, police has to understand what
happened just by listening to stories of accident participants, that not always lead to
correct analysis and respective decisions.
According to [
8
], surveillance video is now the biggest source of Big Data. It is
predicted, that the percentage will increase by 65 percent by 2015. As we are adding
video recordings from users Internet Connected Objects (ICO), using the Internet of
3
Search WWH ::
Custom Search