关键词: 土地利用, 遥感, 时空变化, 瀛湖, 南水北调水源区

Abstract: In order to reveal the land use variation law in Yinghu Lake basin, water sources area of South-to-North Water Diversion, based on Landsat TM/ETM+ and ALOS images, combining supervised classification with manual visual interpretation method, five sets of land use information from 1988 to 2014 were extracted and the spatial and temporal variation characteristics of land utilization and its impacts on water environment were analyzed. The accuracy analysis results showed that: supervised classification combined with manual visual interpretation could extract land use spatial distribution information quickly and accurately, and the accuracy of ALOS image classification could reach more than 98%. However, the image interpretation result of construction land were lower than reality by using moderate or low resolution remote sensing images of Landsat TM/ETM+, because it was difficult to identify decentralized rural residential land in mountain area. The land use classification results show that the land use types in Yinghu Lake basin are interlaced in plaque-like texture due to the influence of geomorphic features of mountain-valley, and construction land distributes in belt form. Grass land, cultivated land and forest land are the main land use types, accounting for 51%, 29% and 18% of the basin area. The annual change rate of construction land was the biggest (more than 22%) since 1988 and the increment mainly came from cultivated land, its contribution proportion was up to 87%.

Key words: land use, remote sensing, spatial-temporal variation, Yinghu Lake, sources area of S-N Water Diversion Project