Этот масштабный открытый набор данных содержит контуры зданий, полученные с помощью спутниковых снимков высокого разрешения (50 см). Он содержит 1,8 млрд обнаруженных зданий в Африке, Латинской Америке, странах Карибского бассейна, Южной и Юго-Восточной Азии. Выводы охватывают территорию площадью 58 млн км².
Для каждого здания в этом наборе данных мы указываем полигон, описывающий его контур на земле, уровень уверенности, показывающий, насколько мы уверены, что это именно здание, и код Plus Code, соответствующий центру здания. Информация о типе здания, его почтовом адресе или какие-либо другие детали, кроме его геометрии, отсутствуют.
Данные о контурах зданий полезны для ряда важных приложений: от оценки численности населения, городского планирования и гуманитарной помощи до изучения окружающей среды и климата. Проект реализуется в Гане, изначально сфокусированный на африканском континенте, а затем включён в исследования Южной Азии, Юго-Восточной Азии, Латинской Америки и Карибского бассейна.
Вывод был сделан в мае 2023 года.
Более подробную информацию можно найти на официальном сайте набора данных Open Buildings.
В. Сирко, С. Кашубин, М. Риттер, А. Аннка, И.С.Э. Бушареб, Ю. Дофин, Д. Кейзерс, М. Нейман, М. Сиссе, Дж. А. Куинн. Обнаружение зданий континентального масштаба по спутниковым снимкам высокого разрешения. arXiv:2107.12283 , 2021.
Этот масштабный открытый набор данных содержит контуры зданий, полученные с помощью спутниковых снимков высокого разрешения (50 см). Он содержит 1,8 млрд обнаруженных зданий в Африке, Латинской Америке, Карибском бассейне, Южной и Юго-Восточной Азии. Выводы охватывают территорию площадью 58 млн км². Для каждого здания в этом наборе данных мы добавляем полигон, описывающий…
[null,null,[],[[["\u003cp\u003eThe Open Buildings dataset provides outlines of 1.8B buildings derived from 50 cm satellite imagery across Africa, Latin America, Caribbean, South Asia, and Southeast Asia.\u003c/p\u003e\n"],["\u003cp\u003eIt includes building footprints, a confidence score, and a Plus Code for each building, covering an area of 58M km².\u003c/p\u003e\n"],["\u003cp\u003eBuilding footprints can be used for various applications, including population estimation, urban planning, and environmental science.\u003c/p\u003e\n"],["\u003cp\u003eThe dataset is available under the CC-BY-4.0 license and is accessible through Google Earth Engine.\u003c/p\u003e\n"],["\u003cp\u003eThe dataset includes building confidence scores, allowing users to filter buildings based on the model's certainty.\u003c/p\u003e\n"]]],[],null,["# Open Buildings V3 Polygons\n\nDataset Availability\n: 2023-05-30T00:00:00Z--2023-05-30T00:00:00Z\n\nDataset Provider\n:\n\n\n [Google Research - Open Buildings](https://sites.research.google/open-buildings/)\n\nTags\n:\n [africa](/earth-engine/datasets/tags/africa) [asia](/earth-engine/datasets/tags/asia) [building](/earth-engine/datasets/tags/building) [built-up](/earth-engine/datasets/tags/built-up) [open-buildings](/earth-engine/datasets/tags/open-buildings) [population](/earth-engine/datasets/tags/population) [south-asia](/earth-engine/datasets/tags/south-asia) [southeast-asia](/earth-engine/datasets/tags/southeast-asia) [table](/earth-engine/datasets/tags/table) \nstructure \n\n#### Description\n\nThis large-scale open dataset consists of outlines of buildings derived\nfrom high-resolution 50 cm satellite imagery. It contains 1.8B building\ndetections in Africa, Latin America, Caribbean, South Asia and Southeast\nAsia. The inference spanned an area of 58M km².\n\nFor each building in this dataset we include the polygon describing its\nfootprint on the ground, a confidence score indicating how sure we are that\nthis is a building, and a [Plus Code](https://plus.codes/) corresponding to\nthe center of the building. There is no information about the type of\nbuilding, its street address, or any details other than its geometry.\n\nBuilding footprints are useful for a range of important applications: from\npopulation estimation, urban planning and humanitarian response to\nenvironmental and climate science. The project is based in Ghana, with an\ninitial focus on the continent of Africa and new updates on South Asia,\nSouth-East Asia, Latin America and the Caribbean.\n\nInference was carried out during May 2023.\n\nFor more details see the official\n[website](https://sites.research.google/open-buildings/) of the Open\nBuildings dataset.\n\n### Table Schema\n\n**Table Schema**\n\n| Name | Type | Description |\n|--------------------|----------|-----------------------------------------------------------------------------|\n| area_in_meters | DOUBLE | Area in square meters of the polygon. |\n| confidence | DOUBLE | Confidence score \\[0.65;1.0\\] assigned by the model. |\n| full_plus_code | STRING | The full [Plus Code](https://plus.codes/) at the building polygon centroid. |\n| longitude_latitude | GEOMETRY | Centroid of the polygon. |\n\n### Terms of Use\n\n**Terms of Use**\n\n[CC-BY-4.0](https://spdx.org/licenses/CC-BY-4.0.html)\n\n### Citations\n\nCitations:\n\n- W. Sirko, S. Kashubin, M. Ritter, A. Annkah, Y.S.E. Bouchareb, Y. Dauphin,\n D. Keysers, M. Neumann, M. Cisse, J.A. Quinn. Continental-scale building\n detection from high resolution satellite imagery.\n [arXiv:2107.12283](https://arxiv.org/abs/2107.12283), 2021.\n\n### Explore with Earth Engine\n\n| **Important:** Earth Engine is a platform for petabyte-scale scientific analysis and visualization of geospatial datasets, both for public benefit and for business and government users. Earth Engine is free to use for research, education, and nonprofit use. To get started, please [register for Earth Engine access.](https://console.cloud.google.com/earth-engine)\n\n### Code Editor (JavaScript)\n\n```javascript\n// Visualization of GOOGLE/Research/open-buildings/v3/polygons.\n\nvar t = ee.FeatureCollection('GOOGLE/Research/open-buildings/v3/polygons');\n\nvar t_065_070 = t.filter('confidence \u003e= 0.65 && confidence \u003c 0.7');\nvar t_070_075 = t.filter('confidence \u003e= 0.7 && confidence \u003c 0.75');\nvar t_gte_075 = t.filter('confidence \u003e= 0.75');\n\nMap.addLayer(t_065_070, {color: 'FF0000'}, 'Buildings confidence [0.65; 0.7)');\nMap.addLayer(t_070_075, {color: 'FFFF00'}, 'Buildings confidence [0.7; 0.75)');\nMap.addLayer(t_gte_075, {color: '00FF00'}, 'Buildings confidence \u003e= 0.75');\nMap.setCenter(3.389, 6.492, 17); // Lagos, Nigeria\nMap.setOptions('SATELLITE');\n```\n[Open in Code Editor](https://code.earthengine.google.com/?scriptPath=Examples:Datasets/GOOGLE/GOOGLE_Research_open-buildings_v3_polygons)\n\n### Visualize as a FeatureView\n\n\nA `FeatureView` is a view-only, accelerated representation of a\n`FeatureCollection`. For more details, visit the\n[`FeatureView` documentation.](/earth-engine/guides/featureview_overview)\n| **Important:** Earth Engine is a platform for petabyte-scale scientific analysis and visualization of geospatial datasets, both for public benefit and for business and government users. Earth Engine is free to use for research, education, and nonprofit use. To get started, please [register for Earth Engine access.](https://console.cloud.google.com/earth-engine)\n\n### Code Editor (JavaScript)\n\n```javascript\nvar fvLayer = ui.Map.FeatureViewLayer(\n 'GOOGLE/Research/open-buildings/v3/polygons_FeatureView');\n\nvar visParams = {\n rules: [\n {\n filter: ee.Filter.expression('confidence \u003e= 0.65 && confidence \u003c 0.7'),\n color: 'FF0000'\n },\n {\n filter: ee.Filter.expression('confidence \u003e= 0.7 && confidence \u003c 0.75'),\n color: 'FFFF00'\n },\n {\n filter: ee.Filter.expression('confidence \u003e= 0.75'),\n color: '00FF00'\n },\n ]\n};\n\nfvLayer.setVisParams(visParams);\nfvLayer.setName('Buildings');\n\nMap.setCenter(3.389, 6.492, 17); // Lagos, Nigeria\nMap.add(fvLayer);\nMap.setOptions('SATELLITE');\n```\n[Open in Code Editor](https://code.earthengine.google.com/?scriptPath=Examples:Datasets/GOOGLE/GOOGLE_Research_open-buildings_v3_polygons_FeatureView) \n[Open Buildings V3 Polygons](/earth-engine/datasets/catalog/GOOGLE_Research_open-buildings_v3_polygons) \nThis large-scale open dataset consists of outlines of buildings derived from high-resolution 50 cm satellite imagery. It contains 1.8B building detections in Africa, Latin America, Caribbean, South Asia and Southeast Asia. The inference spanned an area of 58M km². For each building in this dataset we include the polygon describing ... \nGOOGLE/Research/open-buildings/v3/polygons, africa,asia,building,built-up,open-buildings,population,south-asia,southeast-asia,table \n2023-05-30T00:00:00Z/2023-05-30T00:00:00Z \n-90 -180 90 180 \nGoogle Earth Engine \nhttps://developers.google.com/earth-engine/datasets\n\n- [](https://doi.org/https://sites.research.google/open-buildings/)\n- [](https://doi.org/https://developers.google.com/earth-engine/datasets/catalog/GOOGLE_Research_open-buildings_v3_polygons)"]]
