An Image-Based Urban Flood Detection System using Deep Learning and Conformal Prediction for Edge Deployment

• Guilherme Bisse Alves Rodrigues UNESP
• Arthur Lima Marques Rocha USP
• Jó Ueyama USP
• Caetano Mazzoni Ranieri UNESP

DOI: https://doi.org/10.5753/eniac.2026.30074

Resumo

Urban floods threaten infrastructure and public safety, making reliable early detection systems essential. In this context, this paper presents a solid deep learning pipeline for flood detection using real-world stream images. Due to the underrepresentation of flood images, the proposed architecture was strengthened with a semantic masking data augmentation approach, and conformal prediction was applied to set statistical confidence thresholds for model inference. Using a temporal hold-out strategy, the proposed system achieved a significant F1-score of 78.65% for the critical overflow class, successfully minimizing serious false negatives and demonstrating its operational reliability for real-time monitoring of urban flooding.

Referências

Atok, D. A. M. and Chai, S. S. (2025). Addressing overfitting and overestimation challenges in landslide susceptibility modeling: a case study of penang island, malaysia. Natural Hazards, 121(11):13577–13604.

Balasubramanian, V., Ho, S.-S., and Vovk, V. (2014). Conformal prediction for reliable machine learning: theory, adaptations and applications. Newnes.

Bosquet, B., Cores, D., Seidenari, L., Brea, V. M., Mucientes, M., and Del Bimbo, A. (2023). A full data augmentation pipeline for small object detection based on generative adversarial networks. Pattern Recognition, 133:108998.

Chavula, P., Kayusi, F., Lungu, G., and Uwimbabazi, A. (2025). The current landscape of early warning systems and traditional approaches to disaster detection. LatIA, (3):77.

Cunha, A. d. S. et al. (2023). A gestão de risco para redução de desastres naturais e a construção de cidades resilientes.

Economic News (2026). Desastres naturais no brasil geram r$ 28 bi em perdas. [link]. Acessado em 21/02/2026.

EL HARIRI, A., MOUITI, M., HABIBI, O., and LAZAAR, M. (2023). Improving deep learning performance using sampling techniques for iot imbalanced data. Procedia Computer Science, 224:180–187.

Ghosh, B., Garg, S., and Motagh, M. (2022). Automatic flood detection from sentinel-1 data using deep learning architectures. ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 3:201–208.

Hashi, A. O., Abdirahman, A. A., Elmi, M. A., Hashi, S. Z. M., and Rodriguez, O. E. R. (2021). A real-time flood detection system based on machine learning algorithms with emphasis on deep learning. International Journal of Engineering Trends and Technology, 69(5):249–256.

He, W., Jiang, Z., Xiao, T., Xu, Z., and Li, Y. (2026). A survey on uncertainty quantification methods for deep learning. ACM Computing Surveys, 58(7):1–35.

Howard, A., Sandler, M., Chu, G., Chen, L.-C., Chen, B., Tan, M., Wang, W., Zhu, Y., Pang, R., Vasudevan, V., Le, Q. V., and Adam, H. (2019). Searching for mobilenetv3. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV).

Jain, J., Singh, A., Orlov, N., Huang, Z., Li, J., Walton, S., and Shi, H. (2023). Semask: Semantically masked transformers for semantic segmentation. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) Workshops, pages 752–761.

Kalogiannidis, S., Kalfas, D., Kontsas, S., Papaevangelou, O., and Chatzitheodoridis, F. (2025). Evaluating the effectiveness of early warning systems in reducing loss of life in natural disasters: a case study of greece. Journal of Risk Analysis and Crisis Response, 15(1):33–33.

Karimi, H. and Samavi, R. (2023). Quantifying deep learning model uncertainty in conformal prediction. In Proceedings of the AAAI Symposium Series, volume 1, pages 142–148.

Kebaili, A., Lapuyade-Lahorgue, J., and Ruan, S. (2023). Deep learning approaches for data augmentation in medical imaging: a review. Journal of imaging, 9(4):81.

Kirillov, A., Mintun, E., Ravi, N., Mao, H., Rolland, C., Gustafson, L., Xiao, T., White-head, S., Berg, A. C., Lo, W.-Y., Dollar, P., and Girshick, R. (2023). Segment anything. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pages 4015–4026.

Li, H., Rajbahadur, G. K., Lin, D., Bezemer, C.-P., and Jiang, Z. M. (2024). Keeping deep learning models in check: A history-based approach to mitigate overfitting. IEEE Access, 12:70676–70689.

Li, Y., Matgen, P., and Chini, M. (2026). Quantifying and communicating uncertainty in sar-based flood mapping via density-aware neural networks and conformal risk control. IEEE Transactions on Geoscience and Remote Sensing.

Lin, T.-Y., Goyal, P., Girshick, R., He, K., and Dollar, P. (2017). Focal loss for dense object detection. In Proceedings of the IEEE International Conference on Computer Vision (ICCV).

Musa, A., Kakudi, H. A., Hassan, M., Hamada, M., Umar, U., and Salisu, M. L. (2025). Lightweight deep learning models for edge devices—a survey. International Journal of Computer Information Systems and Industrial Management Applications, 17:18–18.

Pedregosa, F., Varoquaux, G., Gramfort, A., Michel, V., Thirion, B., Grisel, O., Blondel, M., Prettenhofer, P., Weiss, R., Dubourg, V., et al. (2011). Scikit-learn: Machine learning in python. the Journal of machine Learning research, 12:2825–2830.

Ranieri, C. M., Souza, T. L. D. e., Nishijima, M., Krishnamachari, B., and Ueyama, J. (2024). A deep learning workflow enhanced with optical flow fields for flood risk estimation. Applied Intelligence, 54(7):5536–5557.

Roohi, M., Ghafouri, H. R., and Ashrafi, S. M. (2025). Advancing flood disaster management: leveraging deep learning and remote sensing technologies. Acta Geophysica, 73(1):557–575.

Stateczny, A., Praveena, H. D., Krishnappa, R. H., Chythanya, K. R., and Babysarojam, B. B. (2023). Optimized deep learning model for flood detection using satellite images. Remote Sensing, 15(20):5037.

Tellman, B., Sullivan, J. A., Kuhn, C., Kettner, A. J., Doyle, C. S., Brakenridge, G. R., Erickson, T. A., and Slayback, D. A. (2021). Satellite imaging reveals increased proportion of population exposed to floods. Nature, 596(7870):80–86.

Upadhyaya, A. N., Debnath, P., Rani, P., Selvakumar, P., Manjunath, T., and Bhattacharya, S. (2025). Ai-based disaster prediction and early warning systems. In AI and emerging technologies for emergency response and smart cities, pages 55–74. IGI Global Scientific Publishing.

Yang, S., Xiao, W., Zhang, M., Guo, S., Zhao, J., and Shen, F. (2022). Image data augmentation for deep learning: A survey. arXiv preprint arXiv:2204.08610.

Zhan, J. (2024). Mobilenet compression and edge computing strategy for low-latency monitoring. Journal of Computer Science and Software Applications, 4(4).