| Lidar on Infrastructure
Eyes
on the future Machine vision is evolving. Whereas once, getting a true, live
picture of traffic movements was reliant on video, now
technologies such as lidar are opening up new possibilities.
Tom Stone uncovers some recent research projects that are
pushing the boundaries of what machines can see
Today, traffic managers have
a host of advanced machine
vision systems at their
disposal, often deployed at
intersections. Cameras of different
types (enabled by increasingly
advanced software) can perform
a variety of functions, including:
incident detection, to enhance live
traffic management; ANPR/ALPR for
enforcement; and vehicle
classification, to aid with asset
management and roadway design.
But, while traffic cameras continue
to be improved in terms of efficiency
and accuracy, the basic technology
remains unchanged: they all use
video. But now we may be standing
on the brink of a new era of machine
vision for traffic management. One
that does away with video
completely and instead
embraces a new technology,
one commonly associated
with autonomous vehicles.
Tests taking place around
the world suggest that lidar is
the next big thing. In the future road
authorities may have to start looking
at deploying lidar on infrastructure
to support autonomous vehicles.
Lidar is able to create detailed
pictures that will initially benefit
AVs, but ultimately could make roads
safer and more efficient for all users.
We are planning for large scale
deployment of autonomous
shuttles and buses in three of our towns,
for connected use in the early 2020s
Dr Swan Gin Beh, chairman, Economic Development
Board, Singapore
While lidar (light detection and
ranging) isn’t machine vision in the
traditional sense – it will never be
able to read number plates, for
example – using point clouds it can
create such accurate 3D models of the
world it’s no wonder lidar sensors are
often referred to as lidar ‘cameras’.
In one research project, currently
underway at CETRAN (Centre of
January/February 2020 Traffic Technology International 017
www.TrafficTechnologyToday.com
/www.TrafficTechnologyToday.com