基于观测对象的地理空间信息频度统计及可视化

发布时间：2018-07-05 05:28

本文选题：地理空间数据 + 观测对象　；参考：《中国科学院大学(工程管理与信息技术学院)》2013年硕士论文

【摘要】：随着遥感技术不断迅猛的发展,地理空间数据呈爆炸式的增长,如何有效的组织海量的地理空间数据,从宏观的角度掌握不同区域不同时段数据的频度及数据的分布状况,有着很强的应用需求和重要的意义。通过对国内外研究进展的调研与分析发现,地理空间信息统计需要满足特定的需求,目前国内将统计信息与地理空间关联研究应用的较少,没有较为完善的地理空间信息统计分析方法体系。本文尝试以应用需求为指导,以观测对象作为地理空间信息的统计主体,提出了从数据组织到数据频度统计模型研究再到统计结果可视化一套技术方案,主要研究内容包括： (1)设计了基于地理编码框架和观测对象的统一数据组织模型,可以有效地建立地理空间信息数据与观测对象的关联,较好地实现了数据之间的共享利用； (2)在研究了传统的几种频度统计方法的基础之上提出了结合时间的隐含关联关系频度统计方法。通过在协同过滤挖掘隐含关联关系的基础上,结合图模型解释性强的特点,统计出不同观测对象的频度,有利于更好地反映整体趋势； (3)结合现有的可视化技术提出了基于融合的统计结果层级可视化技术,针对海量地理空间数据及大量的观测对象,采用层级聚类的方法对统计结果进行融合,实现了统计数据从宏观到微观、逐步细化的可视化方式。本课题将海量地理空间数据与观测对象建立映射关联关系,通过对该映射关系统计方法的研究,实现地理空间数据的频度统计,并通过多种可视化手段直观高效的对统计结果可视化,目的是提升海量地理空间信息宏观应用及决策的价值。
[Abstract]:With the rapid development of remote sensing technology, the geospatial data is increasing explosively. How to effectively organize the massive geospatial data and master the frequency and distribution of the data in different regions and different periods from the macro perspective. Has the very strong application demand and the important significance. Through the investigation and analysis of domestic and foreign research progress, it is found that geospatial information statistics needs to meet the specific needs. There is no perfect statistical analysis system of geospatial information. Under the guidance of the application requirement and taking the observation object as the statistical subject of geospatial information, this paper proposes a set of technical scheme from data organization to data frequency statistical model research to statistical result visualization. The main research contents are as follows: (1) A unified data organization model based on geo-coding framework and observation objects is designed, which can effectively establish the association between geospatial information data and observation objects. (2) based on the study of several traditional frequency statistics methods, the implicit correlation frequency statistical method combined with time is proposed. On the basis of collaborative filtering mining implicit association relation and the strong explanatory characteristics of graph model, the frequency of different observation objects is calculated, which is helpful to better reflect the overall trend. (3) combined with the existing visualization technology, the hierarchical visualization technology based on fusion is put forward. For the massive geospatial data and a large number of observation objects, the hierarchical clustering method is used to fuse the statistical results. The visualization of statistical data from macro to micro is realized. In this paper, the mapping relation between massive geospatial data and observed objects is established, and the frequency statistics of geospatial data is realized by studying the statistical method of mapping relation. In order to enhance the value of macro application and decision making of massive geospatial information, we visualize the statistical results directly and efficiently by various visualization methods.
【学位授予单位】：中国科学院大学(工程管理与信息技术学院)
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：P208

【参考文献】