Design And Implementation Of A Platform For Traffic Spatial Temporal Data Processing And Interactive Visualization

With the development of society and the acceleration of urbanization,the rapid increase of traffic and travel lead to the increasingly serious condition of urban traffic congestion.With the rapid development of associated technology,more and more traffic data including floating car data are collected and stored.Traffic data has the characteristics of large amount,high dimension and including time and space attributes,which is typical spatio-temporal data and need reasonable and efficient analysis tools to discover the hidden rules behind the data.The rise of intelligent traffic accelerates the development of traffic data research.As an important tool for the analysis of data,Visualization has been applied to the field of traffic data research,which can directly shows the multi-dimensional spatio-temporal trajectory data,explores and finds out the law of urban traffic operation,and provides decision support for urban traffic planning and road construction planning.The existing data visualization solutions in the visual analysis of traffic trajectory data can not take into account technical requirement,generality and visual effects.Based on this,a visualization platform for traffic spatio-temporal data has been designed and implementated.Based on the unified data model,the platform has realized the functions as follows: multi source data integration,conversion,data cleaning and map matching.The design of visualization task logic based on interactive has also completed.The platform has improved the versatility and provided a variety of visualization methods for spatio-temporal data analysis.The main contents of this paper are as follows:The way of data processing has been given,including map data processing,GPS data processing and map matching.Map data is the basis of cartographic visualization.The original map data was processed by ArcGIS Desktop.The map data after processing can meet the needs of the latter analysis tasks.The original GPS data has noise data.Based on the characters of GPS data,this paper has developed the standard representation method of GPS data and the data filtering mechanism,thus the GPS data is expressed by the unified approach and the noise data is filtered and cleaned.There is a deviation between the position determined by positioning system and the actual location.The map matching algorithm can effectively solve this problem.When the traditional map matching method is used to match the low frequency sampling floating car data,the matching accuracy is low.Based on this,an improved floating car map matching algorithm has been put forward: the set of candidate road is determined by the grid partition method based on the dynamic change of the number of sections in the grid and the shortest path and trajectory direction information are added to the process of selecting the optimal matching section.Based on the data processing and the characteristics of traffic trajectory data,requirements analysis has been given including functional requirements and property requirements.Based on the requirements,the task analysis model has been determined,the architecture of the platform has been designed based on layered thought and the four function modules has been designed,including data import module,data processing module,data visualization module and function expansion module.Based on design,each module function of the platform has been implementated,including the function of data input module,data processing module,data visualization module and function.The performance of the system has been tested and analyzed and the results show the platform can meet the design.