Sentinel2 RGB chips over BENELUX with JRC GHSL Population Density 2015 for Learning with Label Proportions
收藏资源简介：
Region of Interest (ROI) is comprised of the Belgium, the Netherlands and Luxembourg We use the communes adminitrative division which is standardized across Europe by EUROSTAT at:<br> https://ec.europa.eu/eurostat/web/gisco/geodata/reference-data/administrative-units-statistical-units<br> This is roughly equivalent to the notion municipalities in most countries. From the link above, communes definition are taken from COMM_RG_01M_2016_4326.shp and country borders<br> are taken from NUTS_RG_01M_2021_3035.shp. <strong>images: Sentinel2 RGB from 2020-01-01 to 2020-31-12</strong><br> filtered out pixels with clouds acoording to QA60 band following the example<br> given in GEE dataset info page at:<br> see https://developers.google.com/earth-engine/datasets/catalog/COPERNICUS_S2_SR_HARMONIZED see also https://github.com/rramosp/geetiles/blob/main/geetiles/defs/sentinel2rgbmedian2020.py <strong>labels: Global Human Settlement Layers, Population Grid 2015</strong> labels range from 0 to 31, with the following meaning:<br> label value original value in GEE dataset<br> 0 0<br> 1 1-10<br> 2 11-20<br> 3 21-30<br> ...<br> 31 >=291 see https://developers.google.com/earth-engine/datasets/catalog/JRC_GHSL_P2016_POP_GPW_GLOBE_V1 see also https://github.com/rramosp/geetiles/blob/main/geetiles/defs/humanpop2015.py<br> <pre><code>_aschips.geojson the image chips geometries along with label proportions for easy visualization with QGIS, GeoPandas, etc. _communes.geojson the communes geometries with their label prortions for easy visualization with QGIS, GeoPandas, etc. splits.csv contains two splits of image chips in train, test, val - with geographical bands at 45° angles in nw-se direction - the same as above reorganized to that all chips within the same commune fall within the same split. data/ a pickle file for each image chip containing a dict with - the 100x100 RGB sentinel 2 chip image - the 100x100 chip level lavels - the label proportions of the chip - the aggregated label proportions of the commune the chip belongs to </code></pre>
感兴趣区域（Region of Interest，ROI）涵盖比利时、荷兰与卢森堡三国。我们采用了由欧盟统计局（EUROSTAT）在欧洲范围内统一规范的市镇（communes）行政区划，相关数据获取地址为：https://ec.europa.eu/eurostat/web/gisco/geodata/reference-data/administrative-units-statistical-units。该行政区划大致对应多数国家的市/镇行政单元。本次数据的市镇边界定义取自COMM_RG_01M_2016_4326.shp文件，国家边界则取自NUTS_RG_01M_2021_3035.shp文件。 **图像数据：2020年1月1日至2020年12月31日的Sentinel-2 RGB影像** 我们依据QA60波段，参考谷歌地球引擎（GEE）数据集说明页面的示例方法，对影像中的云像素进行了过滤处理，相关参考链接如下：https://developers.google.com/earth-engine/datasets/catalog/COPERNICUS_S2_SR_HARMONIZED，亦可参考：https://github.com/rramosp/geetiles/blob/main/geetiles/defs/sentinel2rgbmedian2020.py **标签数据：全球人类居住区图层（Global Human Settlement Layers）2015年人口网格数据** 标签取值范围为0至31，各标签对应原始取值范围如下： | 标签值 | GEE数据集中原始取值范围 | | ------ | ---------------------- | | 0 | 0 | | 1 | 1-10 | | 2 | 11-20 | | 3 | 21-30 | | ... | ... | | 31 | ≥291 | 相关参考链接：https://developers.google.com/earth-engine/datasets/catalog/JRC_GHSL_P2016_POP_GPW_GLOBE_V1，亦可参考：https://github.com/rramosp/geetiles/blob/main/geetiles/defs/humanpop2015.py _aschips.geojson：包含影像切片的几何信息与标签占比，可直接用于QGIS、GeoPandas等工具进行可视化。 _communes.geojson：包含市镇的几何信息与对应标签占比，可直接用于QGIS、GeoPandas等工具进行可视化。 splits.csv：包含影像切片的训练集、测试集、验证集划分方案，划分依据为西北-东南方向45°角的地理带，且同属一个市镇的所有影像切片会被划分至同一子集。 data/目录下包含每个影像切片对应的Pickle文件，每个文件均为一个字典，包含以下内容： - 100×100分辨率的Sentinel-2 RGB影像切片 - 100×100分辨率的切片级标签 - 该切片的标签占比 - 该切片所属市镇的聚合标签占比
