{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/using-convolutional-networks-and-satellite","title":"Using convolutional networks and satellite imagery to identify patterns in urban environments at a large scale","arxiv_id":"1704.02965","date":"2017-04-10","proceeding":null,"authors":["Adrian Albert","Jasleen Kaur","Marta Gonzalez"],"abstract":"Urban planning applications (energy audits, investment, etc.) require an\nunderstanding of built infrastructure and its environment, i.e., both\nlow-level, physical features (amount of vegetation, building area and geometry\netc.), as well as higher-level concepts such as land use classes (which encode\nexpert understanding of socio-economic end uses). This kind of data is\nexpensive and labor-intensive to obtain, which limits its availability\n(particularly in developing countries). We analyze patterns in land use in\nurban neighborhoods using large-scale satellite imagery data (which is\navailable worldwide from third-party providers) and state-of-the-art computer\nvision techniques based on deep convolutional neural networks. For supervision,\ngiven the limited availability of standard benchmarks for remote-sensing data,\nwe obtain ground truth land use class labels carefully sampled from open-source\nsurveys, in particular the Urban Atlas land classification dataset of $20$ land\nuse classes across $~300$ European cities. We use this data to train and\ncompare deep architectures which have recently shown good performance on\nstandard computer vision tasks (image classification and segmentation),\nincluding on geospatial data. Furthermore, we show that the deep\nrepresentations extracted from satellite imagery of urban environments can be\nused to compare neighborhoods across several cities. We make our dataset\navailable for other machine learning researchers to use for remote-sensing\napplications.","url_abs":"http://arxiv.org/abs/1704.02965v2","url_pdf":"http://arxiv.org/pdf/1704.02965v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"using-convolutional-networks-and-satellite","repo_url":"https://github.com/adrianalbert/urban-environments","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}},{"paper_slug":"using-convolutional-networks-and-satellite","repo_url":"https://github.com/annaptasznik/NAIP_PoolDetection","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}},{"paper_slug":"using-convolutional-networks-and-satellite","repo_url":"https://github.com/oliviayong/MUeconProgramming-ML","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"classification","task_name":"General Classification"},{"task_slug":"image-classification","task_name":"Image Classification"},{"task_slug":"image-classification","task_name":"image-classification"}],"methods":[],"datasets_introduced":[{"slug":"urban-environments-dataset","name":"Urban Environments","full_name":""}],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1704.02965","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}