タイトル（掲載誌）Proceedings of the CEReS international symposium = CEReS国際シンポジウム資料集
一般注記type:text
[ABSTRACT]Monitoring the location and distributions of land cover changes is important for environmental protection and management, its change is a key to many diverse applications such as forestry, hydrology, and agriculture. Continues time series MODIS data gives major advancement for land cover related project over large geographic region. This study focuses on the use of coarse spatial resolution MODIS time series data (16-day composite) for Eurasia land cover change detection between the year 2003 and 2008, aim to serve an alarm where rapid land cover conversation can be analyzed with higher resolution remote sensing data and find out better change detection method which suitable for global project. In this study three change detection methods were evaluated: Normalized difference vegetation index image differencing (NDVI), Change vector analysis using Tasseled cap transformation (TCT), Change vector analysis using NDVI and Bare soil index (BI).To setting the threshold for possible changed area in each change detection method several scene from Landsat image and Google earth was used as reference data. Three methods compared each other by the performance on different type of land cover change extraction visually. As a result NDVI image differencing method is suitable for detection of forest clear cut and change vector analysis is good for the detection of burnt area and recovered area. Land cover change detection map was created which shows the follow changes: (1)Tropical forest transformation to agriculture land was mainly detected in south east Asia region;(2)Bumt area and vegetation recovered area were detected in north part of Eurasia;(3)Large vegetation decreasing was detected in middle east region which caused by the changing of the weather condition.
連携機関・データベース国立情報学研究所 : 学術機関リポジトリデータベース（IRDB）（機関リポジトリ）