Font Size: a A A

Research On Land Cover Change In China Based On Remote Sensing Time Series Data

Posted on:2007-10-12Degree:DoctorType:Dissertation
Country:ChinaCandidate:R YuFull Text:PDF
GTID:1100360185478909Subject:Cartography and Geographic Information System
Abstract/Summary:PDF Full Text Request
Land-cover changes play a pivotal role in remote sensing analysis, meteorology and environment change, which is an important study domain in remote sensing science. Also, it is the foundation of many study fields such as carbon cycle research and the climatic change analysis. In the past researches, a number of different approaches to analysis the land cover are being pursued. Many methods were used in some specific study region or the corresponding analysis results were obtained by using different remote sensing data, and the systems analysis techniques which considering the dimension change in space/time are still absent. To investigate the methodology of multi-resource data including remote sensing time series data. Main research topics and initiatives in this thesis include:1. This dissertation investigates issues on methodology and application in land cover change detection and analysis, summarizes its existence question and proposes further research fields such as time series data modeling analysis.2. After comparing results of different methods acquiring land cover information, a biomass model is proposed base on MODIS EVI data.3. Further studies on image merge in northwest, China were carried and the classification accuracy of land cover has been improved obviously, which increases from 81% to 89%.4. Based on vegetation index time series data. A number of systems analysis techniques are carried in several province spatial domains, witch can improve analysis accuracy synthetically.5. Considering the change of sensor or spatial scale, the corresponding influences to vegetation index study are analyzed. In most cases, multi-spectral images are more sensitive to change information and noise.6. By using time series modeling, we propose a new change analysis method based on vegetation index time series data, which acquires a high accuracy simulate value.7. A study on land cover change in China using multi-resource data and integrated techniques was carried; finally the following overall change analysis results are acquired, also, we have studied analysis error in PCA method between different remote sensors.With the research results of this thesis, the following conclusions are reached: In flat study areas which are fit to use linear spectral unmixing, the vegetation coverage may be...
Keywords/Search Tags:land cover, NDVI, linear spectral unmixing, biomass, feature space, time series, principle component analysis, ARMA
PDF Full Text Request
Related items