Analysis And Study Of Traffic Flow Pattern Based On Big Data

With the rapid development of China's economy and the rapid progress of the urbanization process,the urban transportation system faces the increasingly sharp contradiction of the car road.The statistical analysis of urban traffic flow pattern is the most important part of intelligent transportation system,which is of great significance to traffic control decision-making and network resource allocation.In view of the urban road traffic vehicle flow space-time distribution imbalance,unpredictable,easy to cause the problems such as traffic congestion,combined with pattern recognition and data mining technology,put forward the analysis of urban traffic transportation research methods of daily travel mode.Firstly,based on OpenCV cross-platform computer vision library,combined with machine learning algorithm,the vehicle license information image is preprocessed,license plate location and character segmentation.The convolution image recognition algorithm of neural network(CNN)with traditional machine learning algorithm analysis,character recognition of convolution neural network algorithm,greatly improve the computing speed of the character recognition.Secondly by big data mining method,correlation analysis was conducted to treat urban traffic space-time law,combining Apriori algorithm and Hadoop distributed processing framework,design the parallelization Apriori algorithm based on Hadoop,similarity of travel regularity in seasons association mining;A parallel k-means clustering algorithm based on Hadoop was designed to divide regional traffic flow.Taking the sample center day as the core,the data mining is carried out for the morning and evening fluctuations of the traffic in each section of the city,and the daily travel model of urban road traffic vehicles is established based on the data mining results.Finally,the network based on graph theory is abstracted as weighted graph point line,combined with the city traffic patterns and historical travel information,heuristic search in the global scope;according to the dynamic function of the right of way,to update the path information into local live,incremental search,in the local scope for real-time vehicle route guidance update.