Samples and 4-class predictions of agroforest-forest frontiers mapping in Peru

This is a database for the study of agroforest-forest frontiers mapping conducted by Wanting Yang, etc. The data contained in the repository includes two parts. 1. the training samples(polygons and rectagle_roi) annotated with Google Earth as reference. We used Google Earth Pro to make the date of Google Earth imagery match with PlanetScope imagery. 2. the prediction from the 4-class model. Users can visualize the data with QGIS or other GIS software.本数据集由杨婉婷等人为开展农林-森林边界制图研究而构建。 该数据集存储库所包含的数据分为两部分： 1. 训练样本（多边形标注框与矩形感兴趣区域（Rectangle ROI））：以谷歌地球（Google Earth）影像作为参考基准进行标注。我们使用谷歌地球专业版（Google Earth Pro），将谷歌地球影像的成像日期与PlanetScope卫星影像（PlanetScope）进行匹配对齐。 2. 四分类模型的预测结果。 用户可借助QGIS或其他地理信息系统（GIS）软件对该数据进行可视化处理。