Abstract
This study evaluated the performance of the U-Net architecture for identifying informal settlements in high-resolution orthoimagery, using manually annotated masks as reference. Models trained with and without data augmentation were compared. The model with data augmentation achieved better performance in terms of IoU, F1-Score, and Precision, while the model without augmentation obtained higher Recall, highlighting the trade-off between sensitivity and false-positive control. Despite challenges related to small areas and low visual contrast, the results confirm the potential of U-Net for mapping informal settlements.
References
ABASCAL, A. et al. Identifying degrees of deprivation from space using deep learning and morphological spatial analysis of deprived urban areas. Journal of Environmental Management, v. 95. 2022. Available at: https://www.sciencedirect.com/science/article/pii/S0198971522000643. Accessed on: 12 dec. 2023.
ALRASHEEDI, K.; DEWAN, A.; EL-MOWAFY, A. Using local knowledge and remote sensing in the identification of informal settlements in riyadh city, Saudi Arabia. Remote Sensing, v. 15, n. 15, p. 3895, 2021. Available at: https://www.mdpi.com/2072-4292/15/15/3895. Accessed on: 25 sep. 2023.
BUSLAEV, A. et al. Albumentations: fast and flexible image augmentations. arXiv preprint arXiv: 1809.06839. 2018. Available at: https://arxiv.org/pdf/1809.06839. Accessed on: 16 jan. 2024.
CINNAMON, J.; NOTH, T. Spatiotemporal development of informal settlements in cape town, 2000 to 2020: An open data approach. Habitat International, v. 112, 2023. Available at: https://www.sciencedirect.com/science/article/pii/S0197397523000139. Accessed on: 11 apr. 2024.
DATA.RIO. Limite Favelas 2019. 2019. Available at: https://www.data.rio/datasets/limite-favelas-2019/explore. Accessed on: 20 dec. 2023.
DATA.RIO. Limite Áreas de Planejamento (AP). 2023. Available at: https://www.data.rio/datasets/b9e30861acfe4bea947e6278a6b30ce3_1/explore. Accessed on: 20 dec. 2023.
DONG, Shiwei. Spatial Stratification Method for the Sampling Design of LULC Classification Accuracy Assessment: A Case Study in Beijing, China. Remote Sensing, v. 14, n. 4, 2022. Available at: https://www.mdpi.com/2072-4292/14/4/865. Accessed on: 03 feb. 2024.
GHAFFARIAN, S.; EMTEHANI, S. Monitoring urban deprived areas with remote sensing and machine learning in case of disaster recovery. Urban Science, v. 9, n. 4, 2021. Available at: https://www.mdpi.com/2225-1154/9/4/58. Accessed on: 12 dec. 2023.
IBGE. Aglomerados Subnormais 2019: Classificação preliminar e informações de saúde para o enfrentamento à COVID-19. 2020. Available at: https://biblioteca.ibge.gov.br/visualizacao/livros/liv101717_notas_tecnicas.pdf. Accessed on: 06 dec. 2023.
IBGE. Favelas e Comunidades Urbanas: IBGE muda denominação dos aglomerados subnormais. 2024. Available at: https://agenciadenoticias.ibge.gov.br/agencia-noticias/2012-agencia-de-noticias/noticias/38962-favelas-e-comunidades-urbanas-ibge-muda-denominacao-dos-aglomerados-subnormais. Accessed on: 05 nov. 2025.
KEMPER, T. et al. Towards an automated monitoring of human settlements in south africa using high resolution spot satellite imagery. International Society for Photogrammetry and Remote Sensing. v. XL-7/W3, [S.n.], 2015. Available at: https://core.ac.uk/download/pdf/38631179.pdf. Accessed on: 4 mar. 2024
LAKSHMANAN, Valliappa; GÖRNER, Martin; GILLARD, Ryan. Practical machine learning for computer vision. " O'Reilly Media, Inc.", 2021.
LING, Antony. Cortiços eram melhores que as favelas. Caos Planejado, 2018. Available at: https://caosplanejado.com/corticos-eram-melhores-que-favelas/. Accessed on: 23 dec. 2023.
LU, W. et al. A Geoscience-Aware Network (GASlumNet) Combining UNet and ConvNeXt for Slum Mapping. Remote Sensing, v. 16, n. 2, p. 260, 2021. Available at: https://www.mdpi.com/2072-4292/16/2/260. Accessed on: 14 may 2024.
MAIYA, S. R.; BABU, S. C. Slum segmentation and change detection: A deep learning approach. arXiv preprint arXiv: 1811.07896. 2018. Available at: https://arxiv.org/pdf/1811.07896. Accessed on: 13 mar. 2024.
MARINS, Paulo. História da Vida Privada no Brasil: Habitação e Vizinhança: Limites da Privacidade no Surgimento das Metrópoles Brasileiras. 3. ed. São Paulo: Companhia Das Letras, 1998. P. 86.
OLIVEIRA, L. T. et al. Capturing deprived areas using unsupervised machine learning and open data: a case study in São Paulo, Brazil. European Journal of Remote Sensing, 2023. Available at: https://www.tandfonline.com/doi/full/10.1080/22797254.2023.2214690. Accessed on: 5 mar. 2024.
UN BRASIL. ONU-Habitat: população mundial será 68% urbana até 2050. NAÇÕES UNIDAS BRASIL. 2022. Available at: https://brasil.un.org/pt-br/188520-onu-habitat-popula%C3%A7%C3%A3o-mundial-ser%C3%A1-68-urbana-at%C3%A9-2050. Accessed on: 17 sep. 2023.
UN BRASIL. Os objetivos de desenvolvimento sustentável no Brasil. NAÇÕES UNIDAS BRASIL. 2023. Available at: https://brasil.un.org/pt-br/sdgs/11. Accessed on: 17 sep. 2023.
PREFEITURA DA CIDADE DO RIO DE JANEIRO. Conselho Estratégico de Informações da Cidade: Atas de Reuniões. 2012. Available at: https://www.rio.rj.gov.br/documents/91329/1f8a19d9-91d6-430d-81f4-52081055114e. Accessed on: 13 nov. 2025.
PREFEITURA DA CIDADE DO RIO DE JANEIRO. Painel Rio: Conheça mais o rio de hoje para construir o rio de amanhã. 2025. Available at: https://pds-pcrj.hub.arcgis.com/pages/unidades. Accessed on: 12 nov. 2025.
PROVOST, Foster; FAWCETT, Tom. Data Science para Negócios: O que você precisa saber sobre mineração de dados e pensamento analítico de dados. Rio de Janeiro: Alta Books, 2016.
RONNEBERGER, Olaf; FISCHER, Philipp; BROX, Thomas. U-Net: Convolutional Networks for Biomedical Image Segmentation. arXiv preprint arXiv: 1505.04597, 2015. Available at: https://arxiv.org/pdf/1505.04597. Accessed on: 16 jan. 2024.
SILVA, Pedro; BIANCHINI, Zélia; DIAS, Antonio. Amostragem: Teoria e Prática Usando R. Rio de Janeiro: [S.n.], 2023. Available at: https://amostragemcomr.github.io/livro/index.html. Accessed on: 2 feb. 2026.
USGS (United States Geological Survey). The Universal Transverse Mercator (UTM) Grid. [S.l.: s.n.], 2001. P. 2. Documento online. Available at: https://pubs.usgs.gov/fs/2001/0077/report.pdf. Accessed on: 10 jan. 2024.
WURM, M. et al. Semantic segmentation of slums in satellite images using transfer learning on fully convolutional neural networks. ISPRS Journal of Photogrammetry and Remote Sensing, v. 150, 2019. Available at: https://www.sciencedirect.com/science/article/pii/S0924271619300383. Accessed on: 11 dec. 2023
This work is licensed under a Creative Commons Attribution 4.0 International License.
Copyright (c) 2026 Jedielso Sales de Souza, Andrea Diniz da Silva, Ian Monteiro Nunes