Development of a High-resolution Geospatial Dataset for Sidewalk and Flood Analysis using Computer Vision
Resumo
Sidewalk infrastructure plays a vital role in advancing sustainable cities by enhancing mobility, promoting social inclusion, and strengthening urban resilience, particularly in the face of climate-related risks such as flooding. In several developing countries, sidewalks are not only of poor quality but also lack consistent and context-sensitive data that would allow for an adequate assessment of their condition. This paper presents the development of a high-resolution multimodal geospatial dataset aimed at analyzing sidewalk infrastructure in urban environments, with a focus on the borough of São Miguel in São Paulo, Brazil, an area marked by frequent flooding and socio-spatial inequalities. By integrating remote sensing, deep learning, and computer vision techniques, the study supports the adaptation and validation of semantic segmentation tools such as Tile2Net for the Brazilian context. The combination of these methods for sidewalk segmentation shows promising results, with preliminary findings demonstrating the effectiveness of Tile2Net integrated with dilation operation. The proposed approach contributes to scalable and socially equitable solutions for urban infrastructure assessment and inclusive territorial planning, with a particular focus on sidewalk networks in flood-prone areas, making them especially relevant for analyzing potential escape routes during such events.
Referências
K. Lowe, “Environmental justice and pedestrianism: Sidewalk continuity, race, and poverty in new orleans, louisiana,” Transportation Research Record, vol. 2598, no. 2598, pp. 119–123, 2016.
R. J. Lores, “O vexame das calçadas,” [Online]. Available: [link], 2024, accessed: 2025-07-27.
P. J. Gibson and W. E. Marshall, “Disparate approaches to maintaining roads and sidewalks: An interview study of 16 u.s. cities,” Transportation Research Record, vol. 2676, no. 9, pp. 553–567, 2022.
R. Stefanidis and A. Bartzokas-Tsiompras, “Pedestrian accessibility analysis of sidewalk-specific networks: Insights from three latin american central squares,” Sustainability, 2024.
K. A. Shastry, B. Jagadish, V. Bhat, and A. Chowdary, “Automated urban infrastructure classification using aerial imagery,” in 2024 International Conference on Advances in Computing and Information Sciences (IACIS), 2024.
S. Liang, A. Guo, B. Liang, Z. Li, Y. Ding, Y. Wang, and F. Chen, Machine Learning and Computer Vision Applications in Civil Infrastructure Inspection and Monitoring. John Wiley Sons, Ltd, 2024, ch. 4, pp. 59–80.
M. Hosseini, A. Sevtsuk, F. Miranda, R. M. C. Junior, and C. T. Silva, “Mapping the walk: A scalable computer vision approach for generating sidewalk network datasets from aerial imagery,” Computers, Environment and Urban Systems, vol. 101, p. 101950, 2023.
V. Bonets and C. Figueiredo, “Vídeos: Após chuvas, ruas ficam debaixo d’água na zona leste de são paulo,” [Online]. Available: [link], 2025, accessed: 2025-07-20.
Prefeitura de São Paulo, Mapa Digital da Cidade de São Paulo, [link], 2025, accessed: 2025-07-20.
——, São Miguel Paulista, [link], 2025, accessed: 2025-07-19.
MapTiler Team, Coordinate Systems Worldwide, [link], 2025, accessed: 2025-07-20.
European Space Agency (ESA), “Introducing sentinel-2,” [link], 2025, accessed: Jul. 20, 2025.
B.-C. Gao, “NDWI—A normalized difference water index for remote sensing of vegetation liquid water from space,” Remote Sensing of Environment, vol. 58, no. 3, pp. 257–266, 1996.
A. Tao, K. Sapra, and B. Catanzaro, “Hierarchical multi-scale attention for semantic segmentation,” 2020. [Online]. Available: [link]
L. van Oorschot, D. Michailidis, N. IJzerman, S. Jansen, and D. Roijers, “Where (not) to cross the street,” [link], 2024.
M. Saberi, T. Lilasathapornkit, F. Nourmohammadi, A. Emami, and M. Damandeh, “Mapping, modeling, and measuring walking and walkability: The role of ai in enhancing urban pedestrian networks,” [link], 2024.
Y. Zhang, B. Howe, and A. Caspi, “Reliable, routable, and reproducible: Collection of pedestrian pathways at statewide scale,” 2024. [Online]. Available: [link]
Y. Zhang, B. Howe, S. Mehta, N.-J. Bolten, and A. Caspi, “Pathwaybench: Assessing routability of pedestrian pathway networks inferred from multi-city imagery,” 2024. [Online]. Available: [link]
R. I. Sultan, C. Li, H. Zhu, P. Khanduri, M. Brocanelli, and D. Zhu, “GeoSAM: Fine-tuning sam with multi-modal prompts for mobility infrastructure segmentation,” 2024. [Online]. Available: [link]
T. Senlet and A. Elgammal, “Satellite image based precise robot localization on sidewalks,” in 2012 IEEE International Conference on Robotics and Automation. IEEE, 2012, pp. 2647—-2653.
L. da Fontoura Costa and R. M. C. Junior, Shape Analysis and Classification: Theory and Practice. CRC Press, Dec. 2010. [Online]. DOI: 10.1201/9781420037555
D. W. Goodall, “A new similarity index based on probability,” Biometrics, vol. 22, pp. 882—-907, 1966. [Online]. DOI: 10.2307/2528080
R Core Team, R: A Language and Environment for Statistical Computing, R Foundation for Statistical Computing, 2025, (Version 4.5.0) [Computer software]. [Online]. Available: [link]
QGIS Project, QGIS Official Site, 2025, accessed on 2025-07-27. [Online]. Available: [link]
Governo Brasileiro, Instituto Brasileiro de Geografia e Estatística (IBGE), [link], 2025, accessed: 2025-07-20.
