Global Positioning System Reference
In-Depth Information
vision and data fusion techniques (see Section 7.7), it will then be possible to have
an improved Whereness capability.
7.4.4
Vehicle Location
Closed circuit TV cameras (CCTVs) used for traffic surveillance have been in
place for many years. More recently, these have been used to automate the
detection of individual vehicles based on number plate reading, which is possible
by advances in fast computer vision and pattern recognition software. For
example, the London Congestion Charging scheme [5], which covers the whole of
the most central part of London, uses fixed (and some mobile) video surveillance
cameras located at 340 sites. These cameras capture the images of the number
plates of moving vehicles within the charging zone. Drivers have to prepay and
register their vehicle details, which are then entered within a database. Camera
images are processed and the resulting vehicle registration numbers are compared
with the database and defaulters fined. Policing has been transformed by the use
of automatic vehicle detection that is now used for speed enforcement, detective
work, and crime prevention (especially antiterrorism).
Cameras are also used for traffic flow detection. The Trafficmaster [6] system
uses roadside cameras to capture random vehicle numbers that are then read again
at different locations on the network and the vehicle speed and thus the congestion
determined. An RDS-TMC service (see Section 4.1) is offered to motorists
equipped with suitable sat-nav equipment, usually via an integrated product
included within the vehicle when manufactured.
It is probable that there will always be camera surveillance regardless of other
advances in vehicle navigation and automatic driving aids. It seems a lost
opportunity, therefore, that the positioning possibilities derived from the roadside
camera networks are not being integrated more with other positioning systems,
especially those in vehicles based on GNSS.
The ultimate in vehicle positioning is the computer vision systems being used
by mobile robots. Competitions are being held between robotic vehicles that are
programmed to find their way autonomously across wild terrain and urban
environments (mostly for military applications). As the vehicle progresses the
stereoscopic computer vision system can reconstruct the scene as a 3D model and
then steer the vehicle around a clear path. A similar approach is being taken with
robotic rover vehicles exploring the surface of the planet Mars, and others are
planned. Computer vision systems with scene modeling may be used for consumer
driving assistance in the future, although it seems unlikely that the driver would be
entirely robotic. Stereo vision systems are also used in digital navigational road
mapmaking (see Figure 8.1).
