Efficient Search And Statistics Of Massive Trajectory Data Based On Full-text Search Engine

The trajectory data is a collection of continuous trajectory points,which are typically spatial point objects with temporal and spatial properties.In the context of the rapid development of technologies such as the Internet and GPS positioning,the types,scales and speed of trajectory data are rapidly increasing.Massive trajectory data must contain a lot of useful information.Mining and analyzing these information through different techniques and methods is of great significance to the research and application of spatial data.Data query and statistics are the necessary conditions for data mining analysis,and the basic needs of the daily use of Internet products;efficient data query and statistical speed can not only guarantee the timeliness of data analysis,but also help to enhance the user experience of Internet products.Based on the Elasticsearch full-text search engine framework,this paper takes the taxi trajectory data as an example to conduct an in-depth study on the efficient query and statistical analysis of spatial trajectory data.Firstly,it analyzes the advantages and disadvantages of traditional relational database and full-text search engine in large-scale data storage and retrieval.Secondly,it builds Elasticsearch cluster service and tests its query performance.Thirdly,it implements a large number of points data efficient cluster in WebGIS based on Geohash geocoding.Finally,a large-scale trajectory data efficient query statistics system was designed and developed,which verified the feasibility of applying full-text search engine to trajectory data for efficient retrieval.The main results of this research are as follows:1)Exploring the storage plan of the taxi trajectory data in the Elasticsearch cluster and building an Elasticsearch cluster for storing taxi trajectory data.Through the simulation query test of the performance test tool,the cluster has efficient query speed and reliable stability under high concurrent access conditions.2)The back-end implements the related algorithm of efficient dynamic aggregation display of the point data,which solves the problem that the traditional front-end WebGIS framework is difficult to achieve high-efficiency aggregation of large-scale points.Firstly,Geohash is used to mesh the two-dimensional space,grid-based clustering is performed on the data of the taxi starting point,and then the clustering result is again clustered based on density to obtain the final aggregation result.It is verified by examples that the method has better dynamic polymerization efficiency and effect.3)Based on Elasticsearch cluster service and related front-end development framework,a visualization system for efficient query statistics of large-scale trajectory data is designed.The system supports functions such as spatial query,attribute query,aggregate analysis and hotspot analysis.Thereby achieving the combination of full-text search engine and WebGIS application.