关键词: 土地利用, 遥感图像分类, 监督分类, 决策树, 特征指数, 苏南丘陵区

Abstract: Accurate, timely and effective land use information is of great significance to urban development planning as well as resources development and utilization. Jurong river basin was taken as the study area in order to evaluate the applicability of different remote sensing image classification methods in obtaining land cover information. With multi-temporal Landsat 8 OLI images, the land cover of Jurong river basin was obtained by supervised classification methods and decision tree method based on time-series characteristic indices. Comparison between the land cover information obtained from different extraction methods was carried on according to the overall classification accuracy and spatial distribution of ground objects. The results show that the overall classification accuracy and Kappa coefficient of support vector machine was the highest （96.4% and 0.948）, but the decision tree method based on time-series characteristic indices tends to be the closest to reality considering the spatial distribution of different ground objects;Maximum likelihood method and decision tree method were more applicable to land use classification in hilly region of southern Jiangsu province.

Key words: land cover, remote sensing image classification, supervised classification, decision tree, characteristic index, hilly area of southern Jiangsu