Research On Traffic Trajectory Data Processing And Visualization Based On Position And Geometry Relation

The rapid development of society affects the fast-paced growth of urban services,which has caused the problems of high density,heavy pressure and accidents in urban transportation.In order to solve these problems on urban traffic,the related fields of technology are required to be further developed.Correspondingly,the research on the data of the traffic trails is emerging as the filed requires.Trajectory data is a kind of data that includes position,speed,time and other attributes.Because it has the characteristics of high-dimension,large amount data and text as carrier,low efficiency of analysis and serious data loss,as well as no direct visualization are often observed during the analysis process.Therefore,it becomes quite essential to explore a highly efficient and intuitive method in trajectory data analysis.Fortunately,evolution of data mining and visualization technology gradually shows their capacity in solving such problem.Through these new data analysis tools,law behind large amount data could be easily revealed.With visualization technology being applied to display the unrevealing law in different forms,problems in urban transport services can be quickly identified,meanwhile,benefiting the urban traffic management department in decision-making and road planning.However,current trajectory data processing method still could improve its methodology in dealing with efficiency,data loss and effect.There are observed disadvantages such as low efficiency,serious data loss and insignificant effect in the process,due to the random and uneven spatial distribution of the trajectory data.Based on this,the position and geometric relationship of trajectory data will be utilized in this paper.The GPS data of floating vehicles will be researched on through data division and trajectory reconstruction method.Ultimately,visualization of these data could be achieved and this method will later experiment with the taxi data of Zibo City,Shandong Province.The main contents of the thesis are as followed:In view of the fact that most of the dividing methods do not consider the randomness of the spatial distribution of the trajectory data,along with the dividing points are single,which caused the analysis results not as expected,the method of dividing the data points with multi-characteristic trajectory points is proposed.This method overcomes the shortcoming that the division effect is not obvious due to the randomness of space distribution when the trajectory data is divided.Its effect could be easily overserved by the results of the distribution heat map.After the data are divided,this result shows that the method from Author has more obvious effect than the traditional data division method.Nevertheless,after the data is effectively divided,the problem still exists is the data loss situation.In view of the problem that the inverse distance weight interpolation algorithm can't well adapt to the interpolation accuracy when the uneven distribution points are collected,an improved inverse distance weight interpolation method is proposed,combined with the natural neighbor relations and interpolation reconstruction to obtain a new vehicle trajectory.After obtaining a completely new vehicle trajectory data,a visualization method of aggregation and visualization will be applied.Its research outcome will be embedded into the visualization platform of urban traffic data to display the above research intuitively.Based on the characteristics of trajectory data and the relationship between its location and geometry,this paper deals with the data processing from the two dimensions of division and reconstruction,and presents it in a visualized form.It is of great significance to the logic analysis and decision-making of traffic management.