高精度的土地利用制图对国土空间规划和社会经济发展具有重要意义。为快速、准确获取大范围土地利用/覆盖信息，以洞庭湖区为研究对象，以Google Earth Engine (GEE)云平台为支撑，基于Sentinel-2多光谱影像和Sentinel-1 SAR影像，利用Stacking集成学习算法进行土地覆盖类型分类。结果表明，该方法能较好地区分不同土地利用类型，总体精度和Kappa系数分别达到了88.53%和0.85。研究结果以期为大尺度土地利用制图以及国土资源管理与规划提供科学方法和依据。

Highly accurate land-use mapping is of great importance to the territory spatial planning and socio-economic development. In order to obtain large scale land-use/land coverage(LULC) information quickly and accurately, taking Google Earth Engine(GEE) cloud platform as the research platform, taking Dongting Lake as the research object, based on Sentinel-2 multi-spectral images and Sentinel-1 SAR images, we used Stacking integrated learning algorithm to classify land coverage types. The results show that the method could better distinguish different land-use types, and the overall accuracy and Kappa coefficient reach 88.53% and 0.85, respectively. The results of this paper are intended to provide a scientific method for large-scale land-use mapping and territory resource management and planning.