Automatic detection
The responsibility therefore
falls on the traffic technology
industry to give law
enforcement a leg up. Officers
need to be able to spot illegal
cell phone users quickly and at
a distance – with immediate and
indisputable evidence of the
offense. That’s why Roadmetric
has developed the first
commercially-available, videobased,
automated detector for
illegal cell phone use.
Roadmetric Enforcement
Deputy software platform
already uses video analytics to
detect speeding, illegal parking,
illegal use of public transport
| Technology Profile
Above:
Charts showing
police-reported
crashes and total
road fatalities in
the US, between
2007 and 2016
| Free reader
inquiry service
Roadmetric
To learn more about this advertiser, please
visit: www.magupdate.co.uk/ptti
July/August 2019 Traffic Technology International 061
www.TrafficTechnologyToday.com
PHOTOGRAPH: W. ROBERT HOWELL
detecting cell phone users
according to driver posture,
even in video frames where the
phone is not visible at all.
This provides a major
advantage. When the system
detects that a cell phone is
being used, the entire event is
permanently recorded –
including a customizable
amount of footage from before
the detection, thanks to the
cameras which are always
on. The only thing that
authorities need, after all,
is a mere glimpse of the cell
phone in use.
With up to 60 frames per
second, and zoom and video
adjustment settings, authorities
are likely to capture a frame that
can satisfy a traffic court judge.
Every year, road accidents
claim more than a million
lives worldwide. Different
technologies, such as
Roadmetric Enforcement
Deputy, can help to reduce
this number for the better.
Below: The consequences
of using a cellphone while
driving are often serious
lanes, and other common traffic
offenses. Cell phone detection,
though, is a more difficult
problem, and requires the use of
an advanced neural network
that is trained to detect not
only the cell phone itself but the
body postures and behaviors
typical of cell phone users. In
this way, violations can be
detected even if the phone is
partially obscured by the
driver’s head or hands – or by
the driver actively trying to
keep it out of the view of people
outside the car.
Roadmetric has found
that the neural network is
performing beyond expectations,
/ptti
/www.TrafficTechnologyToday.com