Along with the rapid and stable development of social economy,people’s living standard is also improving day by day.Nowadays,the number of private cars has increased greatly.This has led to frequent traffic jams on urban roads.Solving traffic jams is not just a problem for China,but a public problem for the whole world.Through comparative analysis of domestic and foreign research status in the field of traffic congestion,this paper carries out the following research work:(1)Study on traffic congestion recognition method based on Alex Net network model based on static traffic images.First,the Alex Net model was trained with the traffic image data set built in this paper,and the structure of the network model,the learning rate,convolutional core size were adjusted to identify the effect.The identification results of the model were analyzed,and the form of the input image was adjusted accordingly.Finally,the congestion identification model Alex Net-h based on convolutional neural network was obtained.The congestion identification accuracy of the model reached 91.64%,and the average identification time of a single sample was0.03 s.The thesis relied on the index of 90% identification accuracy of the project and the goal of rapid congestion identification.(2)The ant colony algorithm in the path planning algorithm is improved.The congestion coefficient is introduced in the ant colony algorithm from both pheromone initialization and pheromone update,making it suitable for path planning in congested environments.Experiments were designed to simulate the road network to verify the improved algorithm and verify its effectiveness.(3)The urban traffic congestion identification system is designed and built.using the congestion recognition model Alex Net-h obtained based on convolutional neural network as the core algorithm,a traffic platform is designed for two groups of people,namely drivers and traffic management personnel.The system can show the real-time traffic situation and historical traffic situation of the city,and provide road congestion information for drivers and traffic management personnel... |