基于GEE和Sentinel-2影像的临沂市土地利用/覆被分类信息提取

Land use/cover classification based on GEE and landsat images in Linyi City,Shandong Province

土地利用/覆被分类信息是土地资源开发利用、生态环境保护、地质灾害治理等工作的重要基础数据。如何实现区域尺度的土地利用/覆被分类信息的快速准确提取具有实际意义。本文以临沂市为研究区,选用2019年Sentinel-2遥感影像,利用Google Earth Engine(GEE)平台提取光谱特征、植被指数特征、纹理特征、地形特征,基于随机森林算法,开展了临沂市土地利用/覆被信息分类和精度分析。结果表明,采用GEE和随机森林算法模式,能快速实现市域尺度土地利用覆被信息分类,总体精度为83.39%,Kappa系数为0.80,分类质量高,在遥感影像分类应用方面具有优势。

Land use/cover classification information is an important basic data for the development and utilization of land resources,ecological environment protection,geological disaster management and other work.How to extract land use/cover classification information quickly and accurately at regional scale is of practical significance.In this paper,Linyi City was selected as the research area,and the Sentinel-2 remote sensing image in 2019 was selected.Using Google Earth Engine(GEE)platform,the spectral features,vegetation index features,texture features and terrain features were extracted.Based on random forest algorithm,land use/cover classification and accuracy analysis were carried out.The results show that,by using GEE and random forest algorithm model,the urban scale land use/cover classification can be realized quickly.And the overall accuracy is 83.39%,the Kappa coefficient is 0.80,the classification quality is high.It has advantages in the classification applications of remote sensing image.