基于GIS的遥感影像土地利用/土地覆盖信息提取研究——以滇西北香格里拉县为例

在GIS辅助下，基于ETM＋遥感卫星数据，采用分辨率融合提取地物和基于NDVI计算的分区、分层分类以及结合有关地理数据（如DEM、坡度等）基于知识分类方法相结合，通过多步骤信息提取来完成实验区进行土地利用／土地覆盖分类研究，并对分类结果进行精度检验。研究结果表明：该方法不仅消除了单一步骤分类时多类间的相互影响，而且将遥感影像数据与地利信息系统数据有机的结合起来，具有技术操作简单，分类结果理想的特点；对城镇用地、机场用地、水库、湖泊及河流的分类识别精度可达100％，基于知识规则对农村居民地和耕地的提取精度达70％以上，较单一的信息提取方法精度有很大提高。

RESEARCH ON ACQUIRING INFORMATION OF LU/LC BY REMOTE SENSE IMAGE BASED ON GIS——A CASE STUDY OF SHANGRI-LA COUNTY,THE NORTHWESTERN YUNNAN

This paper presents the LUCC of Shangri-la county research by means of Multi-Stage Classification Approach with GIS,based on Land-sat ETM+, and the results show that the extraction techniques are also verified. This method not only removed the influence mutually of several classes with Single-Stage Classification Approach, but also combined the information derived from RS with that derived from GIS, and the procedure of implement is effective; And, the airport, the cities, the reservoir, the lake and the rivers classified recognition precision may reach 100%, withdraws the precision based on the knowledge rule to the countryside inhabitant and the farming to reach above 70%, Comparatively uses the sole information extraction method precision to have a bigger enhancement.