基于Hadoop的GIS空间分析平台关键技术研究
发布时间：2018-12-31 21:32
【摘要】：本文从GIS空间分析和云计算环境下的GIS空间分析的研究背景出发，针对当前GIS数据量大的现状，根据GIS空间分析的原理与方法，结合云计算的Hadoop开源框架，以空间连接操作、缓冲区分析和叠加分析为研究对象，设计了相应的空间分析操作函数，运用MapReduce分布式编程技术，在相应的空间分析操作函数的基础上对这几种空间分析进行并行化设计，并搭建了平台进行实验与分析，主要做了以下工作： （1）分析了当前国内外的GIS空间分析以及GIS空间分析在云计算环境下的研究现状，阐述了空间分析对于GIS的重要性、当前GIS数据量增大对于空间分析的影响以及云计算环境下的Hadoop框架与GIS空间分析相结合的优势。 （2）阐述了Hadoop开源框架对GIS空间数据格式的要求，根据GIS空间数据格式设计了相应的空间数据格式转换函数，实现了空间几何数据格式与Geojson文本数据格式之间的相互转换，分别用于处理输入与输出数据的格式。 （3）为了更有效的进行云计算下的GIS空间分析，针对空间数据和GIS空间分析的特点进行了空间数据的划分以及云计算环境下的GIS空间分析任务调度，实现了基于空间填充曲线的数据划分以及GIS空间分析任务调度的策略，解决了任务调度的负载均衡问题。 （4）与一般的并行GIS空间分析不同的是，本文搭建了云GIS空间分析框架，并针对具体的空间分析设计了相应的空间分析操作流程。 （5）针对具体的空间分析设计了相应的空间操作函数，并结合具体的空间操作函数设计了相应的并行空间分析算法，实现了云计算环境下的空间分析。 （6）搭建了Hadoop云计算平台和配置了Hive数据仓库运行环境，使用了不同数据量的空间数据在单机以及Hadoop集群环境下分别进行了实验，对各种空间分析的实验结果进行了分析和评价，验证了并行算法的优越性，并且通过不同集群规模进行实验，验证了集群性能良好的可扩展性。 本文是在传统的GIS空间分析原理与云计算技术的基础上，，提出了基于Hadoop的GIS空间分析平台设计理念，并设计了并行空间分析算法，提高GIS空间分析的性能。
[Abstract]:Based on the research background of GIS spatial analysis and GIS spatial analysis in cloud computing environment, according to the current situation of large amount of GIS data, according to the principle and method of GIS spatial analysis, combined with the open source Hadoop framework of cloud computing, this paper uses spatial connection operation. Buffer analysis and superposition analysis are used as research objects. The corresponding spatial analysis operands are designed. Based on the corresponding spatial analysis operands, the parallel design of these spatial analysis functions is carried out by using the distributed programming technology of MapReduce. And built a platform for experiments and analysis, mainly do the following work: (1) analyze the current domestic and foreign GIS spatial analysis and GIS spatial analysis in the cloud computing environment, The importance of spatial analysis to GIS, the influence of increasing GIS data amount on spatial analysis and the advantages of combining Hadoop framework with GIS spatial analysis in cloud computing environment are expounded. (2) the requirements of Hadoop open source framework for GIS spatial data format are expounded. According to the GIS spatial data format, the corresponding spatial data format conversion function is designed, and the conversion between spatial geometric data format and Geojson text data format is realized. Format used to process input and output data, respectively. (3) in order to carry out GIS spatial analysis in cloud computing more effectively, the spatial data partition and GIS spatial analysis task scheduling in cloud computing environment are carried out according to the characteristics of spatial data and GIS spatial analysis. The data partition based on the filling curve and the strategy of GIS spatial analysis task scheduling are implemented, and the load balancing problem of task scheduling is solved. (4) different from the general parallel GIS spatial analysis, this paper builds the cloud GIS spatial analysis framework, and designs the corresponding spatial analysis operation flow for the specific spatial analysis. (5) the corresponding spatial operation function is designed for the specific spatial analysis, and the corresponding parallel spatial analysis algorithm is designed in combination with the specific spatial operation function to realize the spatial analysis in the cloud computing environment. (6) build the Hadoop cloud computing platform and configure the Hive data warehouse running environment, and use the spatial data of different data amount to carry on the experiment under the single machine and the Hadoop cluster environment respectively. The experimental results of various spatial analysis are analyzed and evaluated to verify the superiority of the parallel algorithm and the good scalability of the cluster is verified by the experiments of different cluster sizes. Based on the traditional GIS spatial analysis principle and cloud computing technology, this paper puts forward the design concept of GIS spatial analysis platform based on Hadoop, and designs parallel spatial analysis algorithm to improve the performance of GIS spatial analysis.
【学位授予单位】：江西理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：P208
本文编号：2397139
[Abstract]:Based on the research background of GIS spatial analysis and GIS spatial analysis in cloud computing environment, according to the current situation of large amount of GIS data, according to the principle and method of GIS spatial analysis, combined with the open source Hadoop framework of cloud computing, this paper uses spatial connection operation. Buffer analysis and superposition analysis are used as research objects. The corresponding spatial analysis operands are designed. Based on the corresponding spatial analysis operands, the parallel design of these spatial analysis functions is carried out by using the distributed programming technology of MapReduce. And built a platform for experiments and analysis, mainly do the following work: (1) analyze the current domestic and foreign GIS spatial analysis and GIS spatial analysis in the cloud computing environment, The importance of spatial analysis to GIS, the influence of increasing GIS data amount on spatial analysis and the advantages of combining Hadoop framework with GIS spatial analysis in cloud computing environment are expounded. (2) the requirements of Hadoop open source framework for GIS spatial data format are expounded. According to the GIS spatial data format, the corresponding spatial data format conversion function is designed, and the conversion between spatial geometric data format and Geojson text data format is realized. Format used to process input and output data, respectively. (3) in order to carry out GIS spatial analysis in cloud computing more effectively, the spatial data partition and GIS spatial analysis task scheduling in cloud computing environment are carried out according to the characteristics of spatial data and GIS spatial analysis. The data partition based on the filling curve and the strategy of GIS spatial analysis task scheduling are implemented, and the load balancing problem of task scheduling is solved. (4) different from the general parallel GIS spatial analysis, this paper builds the cloud GIS spatial analysis framework, and designs the corresponding spatial analysis operation flow for the specific spatial analysis. (5) the corresponding spatial operation function is designed for the specific spatial analysis, and the corresponding parallel spatial analysis algorithm is designed in combination with the specific spatial operation function to realize the spatial analysis in the cloud computing environment. (6) build the Hadoop cloud computing platform and configure the Hive data warehouse running environment, and use the spatial data of different data amount to carry on the experiment under the single machine and the Hadoop cluster environment respectively. The experimental results of various spatial analysis are analyzed and evaluated to verify the superiority of the parallel algorithm and the good scalability of the cluster is verified by the experiments of different cluster sizes. Based on the traditional GIS spatial analysis principle and cloud computing technology, this paper puts forward the design concept of GIS spatial analysis platform based on Hadoop, and designs parallel spatial analysis algorithm to improve the performance of GIS spatial analysis.
【学位授予单位】：江西理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：P208
【参考文献】
相关期刊论文 前10条
1 祁向前;;GIS空间分析功能在超市选址中的应用[J];测绘科学;2008年06期
2 吴亮;谢忠;陈占龙;马丽娜;;分布式空间分析运算关键技术[J];地球科学(中国地质大学学报);2010年03期
3 方金云,何建邦;网格GIS体系结构及其实现技术[J];地球信息科学;2002年04期
4 刘义;陈荦;景宁;熊伟;;基于R-树索引的Map-Reduce空间连接聚集操作[J];国防科技大学学报;2013年01期
5 贾婷;魏祖宽;唐曙光;金在弘;;一种面向并行空间查询的数据划分方法[J];计算机科学;2010年08期
6 陈全;邓倩妮;;云计算及其关键技术[J];计算机应用;2009年09期
7 陈优良;翁和霞;;GIS线状缓冲区生成算法研究[J];江西理工大学学报;2008年05期
8 韩勇,陈戈,李海涛;基于GIS的城市地下管线空间分析模型的建立与实现[J];中国海洋大学学报(自然科学版);2004年03期
9 罗军舟;金嘉晖;宋爱波;东方;;云计算:体系架构与关键技术[J];通信学报;2011年07期
10 刘伟;杜培军;李永峰;;基于GIS的山西省矿产资源规划环境影响评价[J];生态学报;2014年10期
相关博士学位论文 前1条
1 康俊锋;云计算环境下高分辨率遥感影像存储与高效管理技术研究[D];浙江大学;2011年
本文编号：2397139
本文链接：https://www.wllwen.com/kejilunwen/dizhicehuilunwen/2397139.html
