Research On The Control System Of Signal In Intelligent Traffic Based On Image Processing

With the development of intelligent transportation,intelligent signal lights have become an effective control device for urban vehicles and pedestrians.Most of Chinese signal light control systems use foreign advanced intelligent traffic signal control systems,or improvement based on the traffic characteristics of Chinese road.The system adopts a fixed timing control method according to the road traffic flow information in different time periods.Although it can effectively alleviate road traffic congestion to a certain extent,there are still issues such as uneven distribution of vehicle and pedestrian travel time,long waiting time,and waste of resources caused by fixed timing.This paper conducts research work on image-based intelligent traffic signal control systems.In combination with traffic camera resources,image processing technology is used to process real-time traffic videos of vehicles and pedestrians to obtain real-time vehicle flow information and pedestrian flow information.Based on real-time data of vehicles and pedestrians,fuzzy control technology is used to dynamically allocate the transit time.In this way,real-time data can be grasped in time,intelligent control can be realized,the passing speed of vehicles and pedestrians can be increased,the number of vehicle stops and pedestrian waiting times can be reduced,resources can be saved,and the environment can be protected.The main work of the paper is summarized as follows:(1)System solution construction and hardware design.Through analysis of system requirements and comparison of existing detection methods,a system scheme is constructed that uses image processing technology to realize the real-time detection of traffic video and the regulation of signal time.STM32 as the main controller is used for the hardware design,it is equipped with an image acquisition module,a flow detection module,a peripheral road condition information access module,a traffic signal control module,and a power module to build a system simulation experiment environment.(2)Image preprocessing and moving target detection analysis.First,the pre-processing operations such as grayscale,filtering,binarization,and morphological processing on the image are applied to solve the problems of blur,noise,and uneven brightness caused by the external environment to the video image.Then,an improved moving target detection algorithm is used to extract the foreground area of the target,vehicle targets are detected by using improved concave analysis to segment the sticky vehicles,color features and edge detection to remove vehicle shadows,and pedestrian targets are detected by training pedestrian head and shoulder classifiers.(3)Research on tracking and counting of moving objects.Different counting methods are used for vehicles and pedestrians by analyzing the driving characteristics of vehicles and pedestrians.In the vehicle count,the improved virtual detection coil method is used to achieve accurate counting of vehicles.In the pedestrian count,an improved pedestrian target tracking algorithm is used for target tracking,and an area of interest is set to achieve accurate pedestrian counting.According to experimental verification,the improved algorithms can achieve accurate counting of vehicle and pedestrian targets and reduce counting error.(4)Research on Intelligent Timing Decision.Combining real-time detection of vehicle flow information and pedestrian flow information,intelligent decision fuzzy controllers are designed for vehicles and pedestrians,respectively.The intelligent decision fuzzy controller of the vehicle is designed based on the current road phase traffic flow information and the vehicle carrying capacity of its downstream roads to make judgments.The intelligent decision fuzzy controller of the pedestrian is designed based on the number of pedestrians waiting and the speed of pedestrians to make fuzzy judgments,to realize intelligent real-time adjustment of traffic signals.(5)Software design of intelligent traffic signal control system.According to the characteristics of traffic signal controling and principle of system working,the software flow of the intelligent traffic signal control system is designed.It is verified through simulation experiments that the system can effectively realize the real-time adjustment and control of signal lights at intersection,and realize the orderly passage of vehicles and pedestrians.The intelligent traffic signal control system based on image processing has been simulated and tested,it has completed the identification and counting of vehicles and pedestrians,and the green light time is accurately outputed through fuzzy control,the design goal of real-time traffic signal control is achieved.The use of this system can improve the passing efficiency of vehicles and pedestrians,reduce the invalid waiting time of signal lights and parking times of vehicles,at the same time,it can realize the information sharing of traffic volume of each road,and meet the pedestrian priority of traffic criteria.