Lightweight Geospatial Raster Data Management System

With the rapid development of geographic information industry,computer technology and aerospace technology in recent years,the massive raster map data acquired through remote sensing satellites is accumulating gradually.How to efficiently store and manage these massive raster data is currently the important issue that needs to be resolved in the implementation of WebGIS products.At present,most domestic and foreign companies or scientific research institutes use distributed file system or space expansion engine based on large-scale database to store and manage massive raster data.However,these two methods bring some problems such as complexity of environment building,difficulty of maintenance,convenience of data migration and many other issues when facing the scenarios of surgery examination,mobile emergency platforms in disaster prevention and reduction departments and mobile terminals for browsing map.In view of the above issues,this thesis designs and implements a data management platform for storing and managing geospatial raster data with strong interactivity,high performance and lightweight in a stand-alone usage scenario.The platform includes client,server and data storage layer.It uses Mapbox's MBTiles standard for processing geospatial raster tiles and uses SQLite database as the underlying storage medium for raster tile data.It completes some functionalities such as tile management,metadata management,pyramid mounting,import,export and tile pyramid recycling.The main research contents of this thesis are as follows:(1)For lightweight storage requirement,we analyze the advantages and disadvantages of using file and common database to store geospatial raster data under the traditional stand- alone condition.The open-standard MBTiles hosted by Mapbox Inc.is used to store and access map tiles,standardizedly and efficiently.(2)For the upper slice system application,we code conversion programs for the two forms about file and database,so that all the geospaital raster tiles are stored according to the MBTiles standard and are widely deduped by using the MD5 digest algorithm to avoid extra disk cost.(3)AngularJS technology is used to write the client program of the platform,and the OpenLayers library is used as the browser-side tile access and display framework.The client-side cache and the server-side cache are used to accelerate to access the tile data and the pyramid metadata respectively to enhance usage experience.(4)We design and implement some modules such as user module,pyramid module,tile management module,metadata management module,cache module,justify mounting module,map browsing module and spatial query module.Among them,in the spatial query module,this thesis designs an improved version of the quadtree index structure which ultimately improves the efficiency of spatial point query and rectangular range query.This quadtree index structure reduces the sapce overhead of building quadtree.(5)Functional tests are performed on the server and client,performance tests on reading tile and writing tile are performed by using Apache Benchmark.In the spatial query performance test,the efficiency of using JDBC SQL query mode and quadtree index mode to perform spatial point query and rectangular range query is compared and analyzed.Based on the above research,design and implementation,the functions implemented by the platform have been verified through local tests under laboratory environment and online tests on Alibaba Cloud.The lightweight geospatial raster data management system has been applied to real scientific research project.Practice has shown that it can provide good GIS service assistance for the usage of scenarios such as the field surgical teams,emergency mobile platforms and mobile terminal on map browsing.