基于GeoTrellis的海量栅格数据计算分布式系统建设

Construction of Distributed System for Computing MassiveRaster Data Based on GeoTrellis

为实现海量栅格数据的分布式处理,本文基于Hadoop和Spark分布式计算技术,结合开源框架GeoTrellis地理处理引擎,设计并实现了一个海量栅格数据的分布式计算系统。重点探讨了海量栅格数据分布式存储和计算架构,并基于GeoTrellis设计并实现了该分布式计算系统。最后,以10 m分辨率的全球地表覆盖地图对系统计算性能进行测试,结果表明:基于GeoTrellis开发的分布式计算系统能够有效地提升海量栅格数据的处理能力。

In order to realize the distributed processing of massive raster data,this article is based on Hadoop and Spark distributed computing technology,combined with the open-source framework GeoTrellis geoprocessing engine to design and implement a distributed computing system of massive raster data.It focuses on the distributed storage and computing architecture of massive raster data,designs and implements this distributed computing system based on GeoTrellis.Finally,the system's computing performance was tested with a 10 m resolution global land coverage map.The results show that the distributed computing system based on GeoTrellis can effectively improve the processing capacity of massive raster data.