Full Text PDF
Impact Factor (2025): 6.9
DOI Prefix: 10.47001/IRJIET
Impact Factor (2025): 6.9
DOI Prefix: 10.47001/IRJIET
Vol 10 No 5 (2026): Volume 10, Issue 5, May 2026 | Pages: 642-650
International Research Journal of Innovations in Engineering and Technology
OPEN ACCESS | Research Article | Published Date: 29-05-2026
Flood is one of the natural calamities that presents major threats to human civilization, infrastructure and environment, which requires timely and accurate detection and monitoring to mitigate its impact. Recent development in deep learning has significantly improved semantic segmentation performance. However, conventional convolutional neural networks often struggle to capture large contextual information while retaining smaller spatial details, which are necessary for precise flood water segmentation. This study proposes a Modified U-Net architecture containing dilated convolutional layers to improve flood water segmentation performance. The proposed model includes dilation rates of 2, 4, and 6 within the encoder and bottleneck of U-Net to increase the receptive field without increasing the number of parameters or losing spatial resolution. The proposed models get trained on the training dataset using a combined loss of Binary cross entropy and Dice loss function with AdamW optimizer, validated on a validation dataset and get tested on an unseen separated test dataset using multiple quantitative metrics. Thus, obtained experimental results shows that the Modified U-Net with a dilation rate 2 achieved the improved results overall, with a test Dice coefficient of 0.8839, mean IoU of 0.8331, precision of 0.9013, recall of 0.8709, and accuracy of 0.9367. This improvement shows that an optimal dilation rate balances global context extraction and boundary localization effectively, which is critical for accurate flood water segmentation. The findings of this study highlight the effectiveness of dilated convolutions in improving semantic segmentation performance for flood monitoring applications.
Flood segmentation, U-Net, dilated convolution, semantic segmentation, deep learning.
Nitesh Singh, Prakash Chandra Prasad, & Anku Jaiswal. (2026). Modified U-Net with Dilated Convolution for Flood Water Segmentation. International Research Journal of Innovations in Engineering and Technology - IRJIET, 10(5), 642-650. Article DOI https://doi.org/10.47001/IRJIET/2026.105087
This work is licensed under Creative common Attribution Non Commercial 4.0 Internation Licence
Alfieri, L., Bisselink, B., Dottori, F., Naumann, G., de Roo, A., Salamon, P., ... & Feyen, L. (2017). Global projections of river flood risk in a warmer world. Earth's Future, 5(2), 171-182.
Derrick Bonafilia, Beth Tellman, Tyler Anderson, and Erica Issenberg. Sen1floods11: A georeferenced dataset to train and test deep learning flood algorithms for sentinel-1. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition workshops, pages 210–211, 2020.
Marco Chini, Renaud Hostache, Laura Giustarini, and Patrick Matgen. A hierarchical split-based approach for parametric thresholding of sar images: Flood inundation as a test case. IEEE Transactions on Geoscience and Remote Sensing, 55(12):6975–6988, 2017.
Alex Krizhevsky, Ilya Sutskever, and Geoffrey E Hinton. Imagenet classification with deep convolutional neural networks. Communications of the ACM, 60(6):84–90, 2017.
Olaf Ronneberger, Philipp Fischer, and Thomas Brox. U-net: Convolutional networks for biomedical image segmentation. In International Conference on Medical image computing and computer-assisted intervention, pages 234–241. Springer, 2015.
Behrokh Bahrami and Homayoun Arbabkhah. Enhanced flood detection through precise water segmentation using advanced deep learning models. J. Civ. Eng. Res, 6(1):1–8, 2024.
Jonathan Long, Evan Shelhamer, and Trevor Darrell. Fully convolutional networks for semantic segmentation. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 3431–3440, 2015.
Sanjida Afrin Mou, Tasfia Noor Chowdhury, Adib Ibn Mannan, Sadia Nourin Mim, Lubana Tarannum, Tasrin Noman, and Jamal Uddin Ahamed. Ai driven water segmentation with deep learning models for enhanced flood monitoring. arXiv preprint arXiv:2501.08266, 2025.
Liang-Chieh Chen, Yukun Zhu, George Papandreou, Florian Schroff, and Hartwig Adam. Encoder-decoder with atrous separable convolution for semantic image segmentation. In Proceedings of the European conference on computer vision (ECCV), pages 801–818, 2018.
Panqu Wang, Pengfei Chen, Ye Yuan, Ding Liu, Zehua Huang, Xiaodi Hou, and Garrison Cottrell. Understanding convolution for semantic segmentation. In 2018 IEEE winter conference on applications of computer vision (WACV), pages 1451–1460. IEEE, 2018.
Hengshuang Zhao, Jianping Shi, Xiaojuan Qi, Xiaogang Wang, and Jiaya Jia. Pyramid scene parsing network. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 2881–2890, 2017.
Fisher Yu and Vladlen Koltun. Multi-scale context aggregation by dilated convolutions. arXiv preprint arXiv:1511.07122, 2015.
Furkan Isikdogan, Alan C Bovik, and Paola Passalacqua. Surface water mapping by deep learning. IEEE journal of selected topics in applied earth observations and remote sensing, 10(11):4909–4918, 2017.
Shengyuan Piao and Jiaming Liu. Accuracy improvement of unet based on dilated convolution. In Journal of Physics: Conference Series, volume 1345, page 052066. IOP Publishing, 2019.
Na-Xia Yang, Guo-Fa Li, Ting-Hui Li, Dong-Feng Zhao, and Wei-Wei Gu. An improved deep dilated convolutional neural network for seismic facies interpretation. Petroleum Science, 21(3):1569–1583, 2024.
RJ Pally and S Samadi. Application of image processing and convolutional neural networks for flood image classification and semantic segmentation. Environmental modelling & software, 148:105285, 2022.