基于Landsat多变量时间序列数据的城市变化过程提取与分析

针对城市建成区变化的多样性，以及现有的利用Landsat时间序列进行城市扩展变化的研究只利用单一变量时间序列、并只提取城市扩展（单一变化类型）的现状，本项目将探索利用从Landsat数据得到的多变量时间序列分析和提取多种类型城市变化的新方法，包括构建定量表达城市特征的新的光谱指数；提出针对年际多变量时间序列的分割方法，以及基于时间序列局部模式的城市变化提取方法；利用研究区的提取结果，对多类城市变化的时空变化进行分析。. 项目所取得的成果，将推进多变量时间序列分析方法的研究，有利于更准确地提取城市变化，更全面合理地认识城市化动态过程，为深入分析城市化对全球和区域环境、生态和社会经济发展的影响提供有效的数据和方法。

Considering that changes in urban built-up areas are diverse, and that existing studies on urban land expansion and changes using Landsat time series data only used single-variable time series data to extract urban land expansion, the only urban change type, this project aims to explore new methodology of extraction of multiple urban land change types by using time series of multiple-variables derived from Landsat data. Specifically, new spectral indices quantifying urban features will be first proposed, and new methods of annual multiple-variable (multivariate) time series segmentation and methods of extracting multiple urban land change types using time series local pattern will be then proposed. The spatio-temporal dynamics of multiple urban land changes will be further analyzed using the results obtained from study areas..The results and findings of this study will contribute to studies in multivariate time series data processing and analysis. The study will also benefit to more accurate extraction of urban land change dynamics and to full understanding of urbanization dynamics. Furthermore, the study will provide effective methods and data for analysis of impacts of urbanization on environment, ecology and socio-economic development at global and regional scales.