Video Recognition Of Crossway's Traffic Flow And Signal Lamp Control System

Because of the rapid increasement of the population and vehicle in cities,the city-trunk-line-traffic-control should be accomplished as soon as possible.Effective city-trunk-line-traffic can greatly reduce environmental pollution and traffic costs.Vedio-detection for traffic flow is one of the most advanced technologies for traffic flow detection,which have such features as convenient setting and low cost for maintenance.In the video detection for traffic flow,the paper don't use the traditional technique that subtracts the background image from the traffic flow image.and directly gets the vehicle's edges,on the base of the distinct difference between the vehicles and the background.In the course of enhancing image's edges,the paper don't also make use of the traditional differencial coefficient,and erases the inner pixels in a similar area by "erase-inner-pixels" method.The paper improves the precision of identifying vehicles by "depth-method",and eliminates the error such as identifying the seeper and traffic-signs of the road with the vehicles.Because full-green-lamp-traffic is not possible on the trunk line in most conditions,the paper moves the real crossroad to the perfect crossroad that is closest to the real crossroad.It reduces the time of stopping and increases the traffic capacity of the road.The vehicles' behavior influences directly the traffic efficiency of the road.The paper produces the arithmetic that controls the place and speed of the vehicle on the trunk line. With the arithmetic,the travelling process of vehicles can be stimulated on the road with or without signal lamps.And then,we can studay the influence of different signal control mode on the trunk line traffic.In the period of non-green-signal-lamp,no vehicle can go beyond the stop-line.The paper gives a new concept,the "grouping vehicle",which successfully sovles the problem that the vehicle how to travel safely when it approaches the stop line.