Frequency Statistic And Visualization For Geo-spatial Information Based On Observed Objects

With the rapid development of remote sensing technology, the amount of geo-spatial data has a big boost.Thus, the research on how to organize that huge quantity of data is very important. And how to mine the frequency and distribution of those data in different area and time window in a macro-view is also very meaningful.According to the investigation and analysis for recent research progresses, the statistic of geo-spatial information should meet the special needs normally. However, there are less related researches and applications for statistical and geo-spatial information, neither the comprehensive analysis method systems. So, this dissertation tries to take observed objects as statistical principal part, and proposes a technical framework for data organization, frequency statistical model and results visualization. The whole procedure is under the guidance of application demands.The contents of this dissertation can be listed as follows:(1) A unify data organization model based on both geo-coding framework and observed objects is proposed. This model can build the relationships between geo-spatial information and observed objects effectively, and realize the sharing and utilization between different data.(2) Based on the research of several traditional frequency statistical methods, a new time combined frequency statistical method for implicit relationships is proposed. This method can not only filters and mines the implicit relationships, but also calculates the frequencies for different observed objects by using the ability of graph models, which can reflect the whole trend much better.(3) Based on existed visualization techniques, a hierarchy visualization technique based on statistical fusion results is proposed. For huge quantity of geo-spatial data and lots of observed objects, this technique uses the hierarchical clustering method to fuse statistical results, and realizes the visualization way in which statistical data can be viewed from macro to micro, and be refined step by step.This dissertation builds the mapping relationships between huge quantity of geo-spatial data and observed objects. Based on those mapping relationships, the frequency of geo-spatial data can be calculated, and various types of visualization methods can be performed for statistical results. The works in this dissertation greatly increase the value of macro-applications and decisions of geo-spatial information.