Traffic Lights Intelligent Image Analysis Based On ARM Embedded Control

At present, the traffic light has the several kinds of control mode.The first is the use of fixed traffic light cycle. The second one is the use of variable traffic light cycle, which means that different traffic light cycle is adjusted by the different periods of time,sometimes traffic wardenr changes the traffic light cycle by manual or telecontrol according to the specific situation.The third is the intelligent variable traffic light cycle, which means that computer change the traffic light cycle according to the traffic actual situation.Nowadays,lots of cities in China are using the first control mode, a small part of the cities are using the second, leading to several problems. For example, one situation that one direction of crossroad have no car with green light makes the other direction of the car in a waiting state,because the traffic light is red.And one situation that a direction have more cars than the other direction with green traffic light also make the direction with more cars in a waiting state.These two situations will cause the waste of the time and resources, and environmental pollution problems,because the vehicle should wait for the traffic light without shutting down the engine. The vehicle in idle also consume the fuel.Besides the fuel without completely combustion cause more air pollution.In order to solve this problem, this paper designed an intelligent traffic light control solution, which determine the time of traffic light and automaticlly adjust the traffic light time according to the actual vehicles situation to reduce the unnecessary waiting-time.This solution uses a camera in each direction at crossroad to collect the direction of image,obtains the current direction of the lane occupy data from image processing analysis,transfers and collects data from three directions by nRF905wireless data transmission module to the other directions machine, finally,controls the traffic light according to the lane occupy data of each direction of the vehicle.This paper tries to use lane occupy based on the area to represent the number of vehicles at a certain moment.To compute lane occupy, this paper fitsrt divides lanes manually, then uses the method of background difference and three frame difference to detect vehicles, analysis lane occupy.At last a kind of traffic light control solution is designed to make a decision that how many cycle for next traffic light cycle,when ends the traffic light in advance and adds traffic light time in a certain direction.