Research On Heterogeneous Data Conversion And Map Service Publishing Technology Of Raster Data

In the course of GIS research,Web GIS is a technology that has experienced several generations of development.With its advantages of convenient use,rich content and intuitive use experience,it has been favored by lots of researchers and normal users.The core of Web GIS is map service technology.Although there are many researches on map service,they are still not comprehensive and perfect.In the Age of Big Data,map service faces with problems about increasing data volume,types and formats,which are beyond the capacity of traditional technology.But in contrast,the relevant research has not been promoted in time to solve these problems.Taking the common data model(CDM)data developed by the University Corporation for Atmospheric Research and popular in fields of ocean and meteorology,and raster big data map service architecture as an example,this paper carries on the following two aspects of research:(1)Parallel algorithm of format conversion for common data modelIn order to solve the problem that it is difficult to publich CDM data as a map service,a high-performance algorithm is designed to convert CDM format to general data format combining with Spark parallel framework.The algorithm divides the CDM content into attribute data and spatial data,which are extracted into Json text and Geo TIFF respectively.In order to improve the operation efficiency of the algorithm,some kinds of related are tested,and finally the technical route which can achieve the best effect is selected.Then verification tests are done,which show that comparing to original data,the result of algorithm get an acceptable error,and these two data are basically equal.The algorithm enables CDM data to be published and shared,and fills the data source gap of current map service.(2)Mixed map service architecture based on multi middlewareAiming at the problems that map service architectures in traditional single mode have less function,poor data compatibility and are difficult to deal with large volume data,this paper tests three kinds of existing mainstream map service middleware(Geo Tools,Geo Trellis and Gdal)in three quantitative aspects including tile generation,storage and reading,and other five qualitative evaluations.Finally,this paper combines all middleware and and proposes a mixed map service based on multi middleware.Through this architecture,the map service publishing technology is improved,and the data access ability and big data support ability of map service are improved.This research implements the high-performance format conversion from CDM data to Geo TIFF,constructs a map service architecture with mixed middleware,improves the data compatibility and comprehensive performance of map service technology,and has reference values and practical significance for Web GIS map service architecture facing big data.