Mass Moving Target Data Management And Application Technology Research And Realization

Moving target data is an intersection of geospatial data and streaming data,which not only has characteristics of geospatial data:space,attribute and time,but also has the characteristics of fast,large amount and continuous arrival.To solve the problem of data management,perfect data model and storage mode are needed.In terms of storage,improper selection of data model will lead to storage redundancy and fragmentation,which will greatly reduce the storage efficiency.In the aspect of real-time data analysis,the delay index and accuracy of data processing are very important,and high time delay and error rate will reduce the value of analysis and processing.Aiming at the problem of data model,the paper research on data management domestic and overseas were reviewed and the corresponding analysis,put forward a model based on three-dimensional subdivision and moving target data model(MOST)combined with a data storage model,the model to improve the original 2d H tree,to join the high dimension,get moving target height index value,and according to the index value and the timestamp of data for data storage.Aiming at the problem of data storage,this paper studies several popular data storage methods.The grid data packaging method is studied,the algorithm optimization and parallel acceleration are carried out,and studies the application methods of mass trajectory data storage service in the distributed cluster environment.In view of the application scenario of real-time trajectory data analysis of moving objects,two solutions of real-time analysis are proposed:the five-point method based trajectory prediction method,which can predict the next position point of moving objects more quickly and in good real-time performance.Based on the historical frequency statistical analysis method of Storm,this method analyzes the historical trajectory frequency with high accuracy.The above two methods solve two key problems of real-time analysis:real-time and accurate.This paper designs and realizes the prototype system of mass moving target data management and application and verifies the feasibility of the technology through experiments.The research results of this paper can be applied to the data storage scenario of large-scale streaming data and the real-time online analysis scenario,providing an applicable technical scheme for this scenario.