基于Sentinel-1和Sentinel-2进行的POPCorn可扩展人口地图
简介
Popcorn（人口普查数据）是一种人口测绘方法，旨在解决绘制精确人口地图的难题，尤其是在数据稀缺的地区。Popcorn 仅使用全球免费提供的哨兵 1 号和哨兵 2 号卫星图像以及少量的人口普查总计数，其准确性超过了许多依赖高分辨率建筑物足迹的传统人口测绘方法。例如，Popcorn 利用不到 400 个区域人口普查计数绘制了卢旺达 100 米分辨率的人口地图，在基加利的准确率达到 66%，平均误差仅为每公顷 10 个居民。Popcorn 方法可检索建筑密集区和当地建筑物占用率的清晰地图，为工业仓库等无人居住的建筑密集区的分布提供更多信息。这使得该方法在城市规划和人道主义工作中具有可解释性和实用性。Popcorn 的目标是使获取高分辨率人口地图的途径民主化，让没有资源开展大规模人口普查活动的地区也能获得这些地图。您可以在这里找到论文全文，并在 Popcorn 人口制图项目页面上找到有关模型和文件的更多信息。POPCORN: High-resolution Population Maps Derived from Sentinel-1 and Sentinel-2
| Category | Details |
|---|---|
| Data Inputs | - Satellite imagery from Sentinel-1 and Sentinel-2 |
| - Coarse regional population counts | |
| Key Features | - 100m ground sampling distance (GSD) |
| - Built-up area and building occupancy rate maps | |
| - Scalable and timely mapping for urban planning and humanitarian use | |
| Example Use Case | - Rwanda population mapping: 66% accuracy in Kigali with minimal census data |
代码
var snazzy = require("users/aazuspan/snazzy:styles");
snazzy.addStyle("https://snazzymaps.com/style/15/subtle-grayscale", "Greyscale");
// load the population density
var popDensity = ee.Image("projects/sat-io/open-datasets/POPCORN/POPCORNv1");
// Define the inferno color palette
var infernoPalette = [
'#000004', '#1b0c41', '#4a0c6b', '#781c81', '#a52c7a', '#cf4446',
'#ed721c', '#fb9b06', '#f7d03c', '#fcffa4'
];
// Define visualization parameters.
var visParams = {
min: 0,
max: 4,
palette: infernoPalette,
opacity: 0.8 // 70% transparent
};
// Mask out the lowest values (e.g., less than a certain threshold)
var threshold = 0.08;
var maskedPopDensity = popDensity.updateMask(popDensity.gt(threshold));
// Add the masked population density layer to the map.
Map.addLayer(maskedPopDensity, visParams, 'Population Density');
// Create a legend
var legend = ui.Panel({
style: {
position: 'bottom-right',
padding: '8px 15px',
}
});
// Create legend title
var legendTitle1 = ui.Label({
value: 'POPCORN',
style: {
fontWeight: 'bold',
fontSize: '32px',
margin: '0 0 4px 0',
padding: '0'
}
});
legend.add(legendTitle1);
// Create another legend title
var legendTitle2 = ui.Label({
value: 'Population Density [People/ha]',
style: {
fontWeight: 'bold',
fontSize: '18px',
margin: '0 0 4px 0',
padding: '0'
}
});
legend.add(legendTitle2);
// Create a continuous color legend
var legendImage = ui.Thumbnail({
image: ee.Image.pixelLonLat().select(0),
params: {
bbox: [0, 0, 1, 0.1],
dimensions: '300x15',
format: 'png',
min: 0,
max: 1,
palette: infernoPalette,
},
style: { margin: '0 0 4px 0' },
});
legend.add(legendImage);
// Create labels for min and max values
var minLabel = ui.Label(visParams.min.toString(), { margin: '0 269px 4px 0' });
var maxLabel = ui.Label(visParams.max.toString(), { margin: '0 0 4px 0' });
// Add labels to the legend
var labelsPanel = ui.Panel([minLabel, maxLabel], ui.Panel.Layout.flow('horizontal'));
legend.add(labelsPanel);
Map.setControlVisibility({all: false});
// Add the legend to the map
Map.add(legend);
// Center map for the rwanda/DRC boarder scene
Map.setCenter(29.244453536522037, -1.6857641047022471, 13); // The third parameter is the zoom level.代码链接
https://code.earthengine.google.com/?scriptPath=users/sat-io/awesome-gee-catalog-examples:population-socioeconomics/POPCORN-POPULATION-DENSITY
应用APP
https://ee-nandometzger.projects.earthengine.app/view/popcornv1-rwa
引用
Metzger, Nando, Rodrigo Caye Daudt, Devis Tuia, and Konrad Schindler. "High-resolution population maps derived from Sentinel-1 and Sentinel-2."
Remote Sensing of Environment 314 (2024): 114383.
网址推荐
0代码在线构建地图应用
https://www.mapmost.com/#/?source_inviter=CnVrwIQs
机器学习
https://www.cbedai.net/xg
授权许可
知识共享 署名 4.0 国际 (CC-BY-4.0) 关键词：人口分布图、发展中国家、人口密度、人道主义行动、机器学习模型 由 Metzger 等人提供：Metzger et al 2024 在 GEE 中进行编辑：Metzger et al 2024 和 Samapriya Roy 最后更新：2024-09-08
