Research On Visualization Method Of Road Traffic Trend Based On Trajectory Clustering

With the rapid development of the economy,the increase of vehicles brings great challenges to the management of traffic.The intelligent traffic systems(ITS)aim to effectively integrate advanced information technology,sensor technology and computer technology.They are used in the transportation management system to manage traffic more efficiently.Analysis of mobile vehicle location data acquired by GPS(global positioning system),WIFI,etc.can be used to mine potentially useful information in complex traffic,which is of great significance for guiding road construction,vehicle management and traffic control,but directly visualizing large amounts of data can make visual clutter rather than gain access to important information.The trajectory clustering first classifies the trajectory data according to the similarity of the spatiotemporal attributes,and then obtains the behavior patterns in the trajectory from the clusters.The disorder of the vehicle trajectory data and the current method lack of a more accurate description of the overall trend of the trajectory.This thesis firstly summarizes the trajectory clustering correlation method and traffic data visualization technology;then proposes a cluster method based on multi-representation trajectory;Finally designed the vehicle trajectory visualization system(VTVS).The main contributions of this thesis are as follows:(1)The trajectory clustering algorithm and traffic data visualization are summarized and analyzed.This thesis introduces five general methods in trajectory clustering and analyzes their advantages and disadvantages firstly.Then,the visualization methods of different data attributes are introduced respectively.The visual coding and layout of the mainstream visualization system are used to provide a theoretical basis for the design of the trajectory data visualization system.(2)The clustering method based on multi-representation trajectory is proposed.Based on the traditional density-based trajectory clustering algorithm,the improved trajectory partitioning method and trajectory segment similarity metric function are used to classify the trajectory data,and the proposed clustering-based multi-representation trajectory generation method is used for the clustering results.The experimental results show that the proposed method has strong applicability and feasibility.(3)The visualization system for trajectory data is proposed.The system filters the trajectory according to the user-defined rules firstly;then generates the hot zone through the adaptive hierarchical clustering method;finally,it uses multiple views and interaction schemes to hierarchically display the high-dimensional attributes.The results of the case analysis show that the system has practical value in guiding the planning and construction of urban public transportation facilities.