基于分布式的遥感数据管理系统研究与实现  

Research and implementation of distributed remote sensing data management system

作　　者：赵小明 黄祥志 张丽丽 余涛 ZHAO Xiaoming;HUANG Xiangzhi;ZHANG Lili;YU Tao(Institute of Aerospace Information Innovation,Chinese Academy of Sciences,Beijing 100094,China;College of Resource and Environment,University of Chinese Academy of Sciences,Beijing 100049,China;Langfang Research Institute of Space Information,Langfang 065001,China;North China Institute of Aerospace Engineering,Langfang 065001,China)

机构地区：[1]中国科学院空天信息创新研究院,北京100094 [2]中国科学院大学资源与环境学院,北京100049 [3]中科空间信息廊坊研究院,河北廊坊065001 [4]北华航天工业学院,河北廊坊065001

出　　处：《电子设计工程》2023年第11期1-5,共5页Electronic Design Engineering

基　　金：国家自然科学基金(41907192);民用航天技术预先研究项目(D040401)。

摘　　要：文中基于分布式系统构建了一种集存储、管理和计算为一体的多源异构遥感数据管理系统,系统使用Apache HBase存储标准化后的遥感瓦片对象;使用MongoDB存储瓦片元数据,并根据Z曲线编码进行范围分区,实现负载均衡;使用Apache Spark+GDAL进行栅格计算;利用改进的Z曲线编码进行空间查询。经过原型系统性能测试实验表明,文中查询方法在百万级数据量的窗口查询下是PostgreSQL的空间索引查询性能的4.25倍,计算时使用惰性加载图像数据的策略,计算效率相比GeoTrellis提高了18%。Based on the distributed system,this paper constructs a multi⁃source heterogeneous remote sensing data management system that integrates storage,management and computing.The system uses Apache HBase to store standardized remote sensing tile objects;MongoDB is used to store tile metadata,and performs range partitioning according to Z-curve coding to achieve load balancing;Apache Spark+GDAL is used for grid computing;Improved Z-curve coding is used to do a spatial query.The prototype system performance test experiment shows that the query method in this paper is 4.25 times higher than the PostgreSQL spatial index query performance under the window query with millions of data volume.The calculation performance is 18%higher than GeoTrellis by using the lazy loading strategy of image data.

关 键 词：瓦片数据 分布式存储 分布式计算 Z曲线 

分 类 号：TP98[自动化与计算机技术]