长时间序列土地利用数据时空索引技术研究

摘要

This paper aims to develop an efficient index technology for the change analysis and trend prediction of massive land-use spatio-temporal data, which was collected in the first and second national land survey. The main research way is that on the bases of analyzing grid index and HR index, this paper puts forward suitable mixed spatio-temporal index for the long-time series land-use data in the light of common scenarios about land-use data model. The results include 1) the index technology treats administrative boundaries as spatial grid, then uses the change investigatedyear as time scale to establish HR-tree index on the bases of grid index, and keeps the spatio-temporal objects of status bases and change increment as index nodes, thereby the hybrid spatio-temporal index structure based on grid and HR-tree is formed. 2) the paper describes related algorithms of index tree conformation, index query, index deletion and insertion in detail based on new index structure. The conclusion is that the paper discusses the time complexity with grid index, HR-tree index and the new hybrid index respectively in several typical scenarios cases of the land-use change analysis in Hannan area of Wuhan. It shows that hybrid index method has better query efficiency, and it meets the needs of land-use change spatio-temporal analysis.%研究目的:研究支持一调、二调海量土地利用时空数据变化分析及趋势预测的高效索引技术.研究方法:在分析格网索引、HR索引基础上,针对常用土地利用时空数据模型应用场景,提出了适于长时间序列土地利用数据检索的混合时空的索引方法.研究结果:(1)以行政区边界来划分空间格网,在格网基础上以变更调查年为时间标尺建立对应的HR树索引,将基态现状和变更增量时空对象的索引信息保存为索引结点,建立了基于格网与HR树的混合时空索引结构;(2)基于此索引结构,研究并实现了时空混合索引树生成、索引查询、索引删除、索引新增的具体算法.研究结论:以某县域土地利用变化分析中的几种典型情境为例,对比分析了格网索引、HR索引和混合时空索引检索时间复杂度,结果表明该索引方法具有较好的查询效率,满足土地利用变化分析数据时空查询的需求.