Very high resolution of Jakarta's urban green space data

Very high resolution of Jakarta’s urban green space data was produced using fusion of three different earth observation data (Planet, Sentinel-1 SAR, and Sentinel-2 MSI). Planet data was first transformed into Red Edge Triangulated Vegetation Index (RTVI) and the Red Edge Triangulated Wetness Index (RTWI) spectral indices before classification. Training and testing data were gathered from Google Street View. Two machine learning algorithms were used (Random Forest (RF) and Classification and Regression Tree (CART)) within the Google Earth Engine (GEE) cloud computing platform. To refine the data, Google's Open Building Data was used, the data was accessible through the GEE data catalog.本数据集通过融合三类不同的地球观测数据（Planet、Sentinel-1 SAR、Sentinel-2 MSI），生成了雅加达城区的极高分辨率城市绿地数据集。首先将Planet数据转换为红边三角植被指数（Red Edge Triangulated Vegetation Index）与红边三角湿度指数（Red Edge Triangulated Wetness Index）两类光谱指数，再开展分类工作。训练与测试数据采集自谷歌街景（Google Street View）。研究在谷歌地球引擎（Google Earth Engine, GEE）云计算平台上，采用随机森林（Random Forest, RF）与分类与回归树（Classification and Regression Tree, CART）两种机器学习算法。为精细化该数据集，使用了谷歌开放建筑数据集（Google's Open Building Data），该数据集可通过GEE数据目录获取。