基于多源数据的若尔盖湿地土地利用遥感自动调查

Multi-Source Data-Based Remote Sensing Auto-Survey of Land Use in Ruoergai Wetland

以2004年10月份的TM数据为基础,以相近时段的ETM、SPOT影像和DEM、坡度坡向等数据为参考,在经过遥感图像增强、地形纠正和去云雪等预处理以后,将以上多源数据形成的待分类图像分割成图斑.并在此基础上采用阈值法分层分类和模糊分类相结合的方法,自动提取了若尔盖湿地的土地利用类型,经检验在定性和定量方面均取得了较高的验证精度。将获取的土地利用数据经修改后进行面积统计并与2000年的数据进行比较,结果表明:若尔盖湿地主要地物类别为大片的中、高覆盖度草地,其次为沼泽湿地,有林地面积较小且集中分布在湿地东北部。近5年若尔盖湿地的高覆盖度草地有向中、低覆盖度草地转变的趋势;湿地大面积减少,裸地面积不断增大,若尔盖湿地正在面临着退化的危机。

Based on the TM images of October, 2004, SPOT images and ETM images of similar time periods, and DEM, slope and aspect related data, etc., images-to-classify, formed of the above-mentioned multi-source data, were divided into patches after pre-treatments, like image enhancement, topographical correction and cloud/snow elimination. On such a basis the hie.rarchical threshold classification method combined with the fuzzy classification method was applied to extraction of information about land use types in the Ruoergai Wetland, with results verified to be quite high in accuracy in qualification and quantification. Based on the obtained and modified land use data, the area of each land use type was calculated and compared with the use data of 2000. Results indicated that large tracts of grasslands high and moderate in coverage were the main land feature in the Ruoergai Wetland, followed by wetlands, and then smaller pieces of forestland, scattered mainly in northeastern part of the Ruoergai Wetland. In recent 5 years, the grassland declined in coverage from high to moderate and even low, the area of wetland decreased sharply and the area of bare land increased steadily. As a result, the Ruoergai Wetland is facing the risk of degradation.