基于HBase的森林防火遥感瓦片大数据存储

摘要

Aiming at solving the problem that multi-source,heterogeneous and mass remote sensing data could not be read rapid-ly and considering the characteristics of forest fire data and remote sensing pyramids,we introduced and applied the Hilbert curve coding to improve index structure,proposing an HBase-based big data storage structure and a parallel storage struc-ture for forest fire remote sensing data.By experimental analysis,the improved index structure considerably accelerates the reading speed,and the parallel storage structure accelerates data storage.Applying this to Daxing'an Mountains suddens forest fire in Mongolia,our work enables numerous forest fire images to be storage and managed more efficiently,and thus has great potential as a widely used practice.%针对目前多源、异构、海量遥感数据难以快速读取的情况,结合森林防火遥感影像数据和遥感影像金字塔的特点,应用Hilbert曲线改进索引结构,提出了基于HBase的森林防火遥感瓦片大数据存储结构;设计了森林防火遥感瓦片数据并行化存储架构,并应用于内蒙古辖区大兴安岭突发森林火灾辅助决策中.通过实验分析,验证了改进的索引结构在读取速度上明显优于改进之前,并行化存储加快了数据的存储.研究成果可满足对海量森林防火影像高效存储、管理的需求,具有重要的实用价值.