|
The traffic congestions at signalized intersections are
recurrent; and it is the jammed condition that makes it
difficult to separate one car from another. In order to
solve this kind of problem, we collect the traffic flow data in
an unconventional way, describing the characteristics of the
traffic flows by analyzing the time-space diagram of
platoons. In the aspect of photographing traffic flows, this
study solves the problem of non-vertical angle photos. We
firstly calibrate the Bleyl coordinate transfer formula
with multi-points. In vehicle recognition, this study
develops the "twice histogram method" to obtain the traffic
image background from which the passing of vehicles is
recognized. Another development about vehicle detection
is "the vehicle-path pseudo-line detector", the so-called
"line & curve detector", for analyzing the time-spacing of
vehicular flows at different lane/road situations. For
traffic parameters, we can get the time-space diagram
every one-thirtieth second and obtain the vehicular flow
shock waves at signalized intersections. It is suitable
for analyzing many kinds of straight and turning paths
of traffic flows. It can also help us collect the parameters of
discharge headway, saturation flow rate, time-space diagram,
lane occupancy, queuing length, delay, turning rate, and
platoon shock wave. Two important conclusions have been
reached in this study: 1.The twice histogram method can
separate the traffic background image from congested vehicle
flow. 2.Vehicle-path pseudo-line detectors provide a
convenient and accurate analysis method for describing
the traffic flows at congested, signalized intersections.
It''s indeed a progress in image processing and in the traffic
engineering field.
|