加拿大农田样本数据集(The Canadian Cropland)

该存储库包含使用新型基于斑块的数据集加拿大农田数据集的说明和代码，该数据集基于Eurosat 数据集。它是使用从 Sentinel-2 检索到的加拿大农业农田光学卫星图像编制的。使用谷歌地球引擎(GEE) 提取了 5 个月（6 月至 10 月）和 4 年（2017-2020 年）间 10 种主要作物类型的共 78,536 张高分辨率地理参考图像，并自动用加拿大作物清单进行标记。每张图像包含 12 个主要光谱波段以及与植被指数（GNDVI、NDVI、NDVI45、OSAVI 和 PSRI）相对应的一系列波段。图像是使用 6,633 个加拿大农田地理点的列表收集的。

This repository contains documentation and code for the novel patch-based Canadian Farmland Dataset, which is derived from the Eurosat dataset. This dataset is compiled using optical satellite imagery of Canadian agricultural farmlands retrieved from Sentinel-2. A total of 78,536 high-resolution georeferenced images covering 10 major crop types were extracted via the Google Earth Engine (GEE) over a 5-month period (June to October) across 4 years (2017–2020), and automatically annotated using the Canadian Crop Inventory. Each image includes 12 core spectral bands, alongside a series of bands corresponding to vegetation indices such as GNDVI, NDVI, NDVI45, OSAVI, and PSRI. The dataset was collected using a list of 6,633 georeferenced points corresponding to Canadian farmlands.