湿原植生分類図作成に有効となる観測時期の選択

抄録

Recent study has pointed out that the wetlands are the major emission source of methane, which is one of the green house gases. To produce the wetland vegetation distribution map with high spatial resolution satellite images are urgent necessity for estimating the amount of methane emission. As an initial classification step, optimal observing seasons and combinations must be selected to improve the accuracy of classification and reducing the data volume. In this study, effective combinations of observing season and band number were selected to produce a wetland distribution map with multispectral and panchromatic satellite sensor data. In the process of determine, test data were classified by maximum likelihood method by changing every combinations of spectral bands from every 6 season's and combined two seasonal data. The best classification accuracy was attained by using 10 bands of combined TM data from July and August which are middle and final growth stages of wetland vegetation.