2692-4048ISSN (Online)
2770-0070ISSN (Print)
Volume 21 , Issue 2 , PP: 336-352, 2026 | Cite this article as | XML | Html | PDF | Full Length Article
Sanket S Kulkarni 1 * , Ansuman Mahapatra 2
Doi: https://doi.org/10.54216/FPA.210221
Floods are among the most devastating natural disasters, causing widespread damage to infrastructure, homes, and human lives. Rapid assessment of flood severity is critical for effective disaster response and resource allocation. This study explores several deep learning approaches for flood water level classification using UAV imagery. A curated dataset of 2,000 UAV images from diverse regions, including India, the United States, and Brazil, was developed and augmented to improve generalization. Multiple architectures were evaluated, including pre-trained CNNs, ResNet50v2, MobileNetv2, Vision Transformers, and Swin Transformers, with and without the Convolutional Block Attention Module (CBAM) and adaptive learning strategies. Experimental results reveal that integrating Vision Transformers with CBAM achieves the highest classification accuracy of 90.6%, while a hybrid CNN–Vision Transformer model further improves performance to 92.3%. These findings highlight the potential of attention-based hybrid models for precise flood severity mapping. The proposed framework can aid rescue teams and disaster management authorities by prioritizing high-risk areas, enabling faster response and optimized allocation of resources during emergency operations.
Flood water level , Pre-trained CNN , CBAM , Vision Transformer , Swin transformer , Hybrid vision transformer
[1] United Nations Office for Disaster Risk Reduction, "Heavy Floods Widespread Across Asia," UNDRR News. https://www.undrr.org/news/heavy-floods-widespread-across-asia.
[2] Betterle and P. Salamon, "Water depth estimate and flood extent enhancement for satellite-based inundation maps," Nat. Hazards Earth Syst. Sci., vol. 24, no. 8, pp. 2817–2836, 2024, doi: 10.5194/nhess-24-2817-2024.
[3] Chamatidis, D. Istrati, and N. D. Lagaros, "Vision Transformer for Flood Detection Using Satellite Images from Sentinel-1 and Sentinel-2," Water, vol. 16, no. 12, p. 1670, 2024, doi: 10.3390/w16121670.
[4] G. K. Wedajo, T. D. Lemma, T. Fufa, and P. Gamba, "Integrating Satellite Images and Machine Learning for Flood Prediction and Susceptibility Mapping for the Case of Amibara, Awash Basin, Ethiopia," Remote Sens., vol. 16, no. 12, p. 2163, 2024, doi: 10.3390/rs16122163.
[5] Z. Li and I. Demir, "U-net-based semantic classification for flood extent extraction using SAR imagery and GEE platform: A case study for 2019 central US flooding," Sci. Total Environ., vol. 869, p. 161757, 2023, doi: 10.1016/j.scitotenv.2023.161757.
[6] W. Li, D. Li, and Z. N. Fang, "Intercomparison of automated near-real-time flood mapping algorithms using satellite data and DEM-Based methods: a case study of 2022 Madagascar flood," Hydrology, vol. 10, no. 1, p. 17, 2023, doi: 10.3390/hydrology10010017.
[7] Elkhrachy, "Flash flood water depth estimation using SAR images, digital elevation models, and machine learning algorithms," Remote Sens., vol. 14, no. 3, p. 440, 2022, doi: 10.3390/rs14030440.
[8] H. Hosseiny, "A deep learning model for predicting river flood depth and extent," Environ. Model. Softw., vol. 145, p. 105186, 2021, doi: 10.1016/j.envsoft.2021.105186.
[9] D. Bonafilia, B. Tellman, T. Anderson, and E. Issenberg, "Sen1Floods11: A georeferenced dataset to train and test deep learning flood algorithms for sentinel-1," in Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit. Workshops, 2020, pp. 210–211, doi: 10.1109/CVPRW50498.2020.00113.
[10] S. Cohen et al., "Estimating floodwater depths from flood inundation maps and topography," JAWRA J. Amer. Water Resour. Assoc., vol. 54, no. 4, pp. 847–858, 2018, doi: 10.1111/1752-1688.12609.
[11] F. Cian, M. Marconcini, P. Ceccato, and C. Giupponi, "Flood depth estimation by means of high-resolution SAR images and lidar data," Nat. Hazards Earth Syst. Sci., vol. 18, no. 11, pp. 3063–3084, 2018, doi: 10.5194/nhess-18-3063-2018.
[12] Stateczny, H. D. Praveena, R. H. Krishnappa, K. R. Chythanya, and B. B. Babysarojam, "Optimized Deep Learning Model for Flood Detection Using Satellite Images," Remote Sens., vol. 15, no. 20, p. 5037, 2023, doi: 10.3390/rs15205037.
[13] Q. C. Le, M. Q. Le, M. K. Tran, N. Q. Le, and M. T. Tran, "FL-Former: Flood Level Estimation with Vision Transformer for Images from Cameras in Urban Areas," in Int. Conf. Multimedia Modeling, Cham: Springer, 2023, pp. 447–459, doi: 10.1007/978-3-031-27077-2_35.
[14] N. Anusha and B. Bharathi, "Flood detection and flood mapping using multi-temporal synthetic aperture radar and optical data," Egypt. J. Remote Sens. Space Sci., vol. 23, no. 2, pp. 207–219, 2020, doi: 10.1016/j.ejrs.2019.01.001.
[15] K. Tanaka, Y. Fujihara, K. Hoshikawa, and H. Fujii, "Development of a flood water level estimation method using satellite images and a digital elevation model for the Mekong floodplain," Hydrol. Sci. J., vol. 64, no. 2, pp. 241–253, 2019, doi: 10.1080/02626667.2019.1578463.
[16] P. Zhong, Y. Liu, H. Zheng, and J. Zhao, "Detection of urban flood inundation from traffic images using deep learning methods," Water Resour. Manag., vol. 38, no. 1, pp. 287–301, 2024, doi: 10.1007/s11269-023-03669-9.
[17] J. Wan et al., "Automatic detection of urban flood level with YOLOv8 using flooded vehicle dataset," J. Hydrol., vol. 639, p. 131625, 2024, doi: 10.1016/j.jhydrol.2024.131625.
[18] M. Ranieri, T. L. de Souza, M. Nishijima, B. Krishnamachari, and J. Ueyama, "A deep learning workflow enhanced with optical flow fields for flood risk estimation," Appl. Intell., vol. 54, pp. 5536–5557, 2024, doi: 10.1007/s10489-024-05466-2.
[19] K. J. Wienhold, D. Li, W. Li, and Z. N. Fang, "Flood Inundation and Depth Mapping Using Unmanned Aerial Vehicles Combined with High-Resolution Multispectral Imagery," Hydrology, vol. 10, no. 8, p. 158, 2023, doi: 10.3390/hydrology10080158.
[20] G. Popandopulo et al., "Flood extent and volume estimation using remote sensing data," Remote Sens., vol. 15, no. 18, p. 4463, 2023, doi: 10.3390/rs15184463.
[21] H. S. Munawar, F. Ullah, S. Qayyum, and A. Heravi, "Application of deep learning on uav-based aerial images for flood detection," Smart Cities, vol. 4, no. 3, pp. 1220–1242, 2021, doi: 10.3390/smartcities4030065.
[22] J. Strebl et al., "Flood level estimation from social media images," in CEUR Workshop Proc., vol. 2670, MediaEval, pp. 1–3, 2019.
[23] A. Kharazi and A. H. Behzadan, "Flood depth mapping in street photos with image processing and deep neural networks," Comput., Environ. Urban Syst., vol. 88, p. 101628, 2021, doi: 10.1016/j.compenvurbsys.2021.101628.
[24] J. L. Gan and W. Zailah, "Water level classification for flood monitoring system using convolutional neural network," in Proc. 11th Natl. Tech. Semin. Unmanned Syst. Technol. 2019: NUSYS'19, Singapore: Springer, 2021, pp. 299–318, doi: 10.1007/978-981-15-5281-6_21.
[25] Y. Feng, C. Brenner, and M. Sester, "Flood severity mapping from Volunteered Geographic Information by interpreting water level from images containing people: A case study of Hurricane Harvey," ISPRS J. Photogramm. Remote Sens., vol. 169, pp. 301–319, 2020, doi: 10.1016/j.isprsjprs.2020.09.011.
[26] K. A. C. Quan et al., "Flood level prediction via human pose estimation from social media images," in Proc. 2020 Int. Conf. Multimedia Retrieval, 2020, pp. 479–485, doi: 10.1145/3372278.3390704.
[27] Y. T. Lin, M. D. Yang, J. Y. Han, Y. F. Su, and J. H. Jang, "Quantifying flood water levels using image-based volunteered geographic information," Remote Sens., vol. 12, no. 4, p. 706, 2020, doi: 10.3390/rs12040706.
[28] H. Rizk, Y. Nishimur, H. Yamaguchi, and T. Higashino, "Drone-based water level detection in flood disasters," Int. J. Environ. Res. Public Health, vol. 19, no. 1, p. 237, 2021, doi: 10.3390/ijerph19010237.
[29] Y. Liang, X. Li, B. Tsai, Q. Chen, and N. Jafari, "V-FloodNet: A video segmentation system for urban flood detection and quantification," Environ. Model. Softw., vol. 160, p. 105586, 2023, doi: 10.1016/j.envsoft.2022.105586.
[30] P. Chaudhary, S. D'Aronco, M. Moy de Vitry, J. P. Leitão, and J. D. Wegner, "Flood-water level estimation from social media images," ISPRS Ann. Photogramm., Remote Sens. Spatial Inf. Sci., vol. 4, no. 2/W5, pp. 5–12, 2019.
[31] Dosovitskiy, "An image is worth 16x16 words: Transformers for image recognition at scale," arXiv: 2010.11929, 2020.
[32] Z. Liu et al., "Swin transformer: Hierarchical vision transformer using shifted windows," in Proc. IEEE/CVF Int. Conf. Comput. Vis., 2021, pp. 10012–10022, doi: 10.1109/ICCV48922.2021.00986.
[33] R. J. Pally and S. Samadi, "Application of image processing and convolutional neural networks for flood image classification and semantic segmentation," Environ. Model. Softw., vol. 148, p. 105285, 2022, doi: 10.1016/j.envsoft.2021.105285.
[34] S. Woo, J. Park, J. Y. Lee, and I. S. Kweon, "Cbam: Convolutional block attention module," in Proc. Eur. Conf. Comput. Vis. (ECCV), 2018, pp. 3–19, doi: 10.1007/978-3-030-01234-2_1.
[35] S. S. Kulkarni and A. Mahapatra, "Improving Satellite Flood Image Classification Using Attention-Based CNN and Transformer Models," Int. J. Adv. Comput. Sci. Appl., vol. 16, no. 3, 2025, doi: 10.14569/IJACSA.2025.01603101.
[36] S. S. Kulkarni and A. Mahapatra, "Post flood assessment using deep learning techniques," in AIP Conf. Proc., vol. 2917, no. 1, p. 050001, AIP Publishing, 2023, doi: 10.1063/5.0175612.
