The Research Of The Intelligent Traffic Control System Based On Digital Image Processing

In today’s society, the car has become people’s one of the necessary traffic tools, at the same time cars brought convenient transportation, because of increasing cars, traffic jams has become more universal. Therefore, we need to establish a rational intelligent traffic control system. This paper combine image processing method and computer vision, used in intelligent transportation system control.Traffic control should be used in different weather, environment, lighting conditions in normal effective work, therefore this paper will be divided into day and night vehicle queue detection. Then through the improvement of expansion window algorithm and three dimensional space to two dimensional image mapping method measured respectively in the night, vehicle queue length. Then according to the vehicle length has formulated the traffic lights long stay. So as to achieve the purpose of intelligent control traffic lights.First of all, for the day vehicle queue detection, the paper detect moving vehicles based on three frame differential method, and detect existing vehicles based on robust to noise type morphologic edge detection. Second, since the the difference of day and night environment is very big, so we are not able to use the day case detection method,this paper use light as a symbol of night vehicles to detect queuing vehicles. Finally this paper use improved expansion window algorithm to measure the length of the queuing vehicles during the day and use three dimensional space to two dimensional image mapping method to detect night queuing vehicles length, and use the KNN algorithm to identify day and night, at last this paper use the measured the vehicle queue length to formulate corresponding traffic light time. So as to achieve the purpose of intelligent traffic control.The experimental results showed that in different weather environment and illumination brightness cases, this method can be accurate and effective detection vehicle queue length, so that can control traffic lights more reasonable.