Evaluating Deep Segmentation Models for Wildfire Mapping Using Sentinel-2 Satellite Imagery

Victor Hugo Lara Souza; Rodrigo Smarzaro; Leandro Henrique Furtado Pinto Silva; João Fernando Mari

doi:10.22456/2175-2745.150922

Authors

Victor Hugo Lara Souza Universidade Federal de Viçosa (UFV) https://orcid.org/0009-0002-6573-9159
Rodrigo Smarzaro Universidade Federal de Viçosa (UFV) https://orcid.org/0000-0002-4783-8383
Leandro Henrique Furtado Pinto Silva Universidade Federal de Viçosa (UFV) https://orcid.org/0000-0002-5765-5206
João Fernando Mari Universidade Federal de Viçosa (UFV) https://orcid.org/0000-0001-8271-3802

DOI:

https://doi.org/10.22456/2175-2745.150922

Keywords:

wildfires, semantic segmentation, deep learning, data augmentation

Abstract

Wildfires pose serious threats to ecosystems, human health, and urban infrastructure, making accurate and efficient mapping of burned areas essential for mitigation and recovery. In this context, the use of computer vision techniques has shown highly effective results for rapid and accurate delineation of burned areas. However, these techniques require considerable computational cost, and their performance is sensitive to model selection and hyperparameter adjustments. Therefore, this work evaluates deep learning models for the semantic segmentation of regions affected by wildfires using multispectral images from the CEMS-Wildfire dataset, obtained from the Sentinel-2 satellite. We evaluated the combination of three traditional architectures: U-Net, Feature Pyramid Network (FPN), and SegFormer, combined with different encoders (ResNet-50, EfficientNet B2, and MiT-B0), under different data augmentation strategies (none, mild, strong) and loss functions (Binary cross-entropy, Dice loss, and hybrid BCE + Dice loss). U-Net achieved the best performance with the MiT B0 encoder, trained using Dice loss and moderate augmentation, resulting in an IoU of 0.7790 and an F1-score of 0.8758.

Downloads

Download data is not yet available.

References

[1] REQUIA, W. J. Fires in Brazil: health crises and the failure of government action. The Lancet Regional Health–Americas, Elsevier, v. 39, 2024.

[2] MAPBIOMAS. MapBiomas Brasil. 2025. https://plataforma.brasil.mapbiomas.org/monitor-do-fogo. Accessed: 2025-09-01.

[3] PHIRI, D. et al. Sentinel-2 data for land cover/use mapping: A review. Remote Sensing, MDPI, v. 12, n. 14, p. 2291, 2020.

[4] SISMANIS, M. et al. An unsupervised burned area mapping approach using Sentinel-2 images. Land, MDPI, v. 12, n. 2, p. 379, 2023.

[5] SILVA, L. H. F. P. et al. Cloud removal with compact diffusion models: A residual block-based approach. Remote Sensing Applications: Society and Environment, Elsevier, p. 101680, 2025.

[6] KHRYASHCHEV, V.; LARIONOV, R. Wildfire segmentation on satellite images using deep learning. In: IEEE. 2020 Moscow Workshop on Electronic and Networking Technologies (MWENT). [S.l.], 2020. p. 1–5.

[7] BRAND, A.; MANANDHAR, A. Semantic segmentation of burned areas in satellite images using a U-Net-based convolutional neural network. The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Copernicus GmbH, v. 43, p. 47–53, 2021.

[8] ARNAUDO, E. et al. Robust burned area delineation through multitask learning. In: SPRINGER. Joint European Conference on Machine Learning and Knowledge Discovery in Databases. [S.l.], 2023. p. 436–447.

[9] IAKUBOVSKII, P. Segmentation Models PyTorch. [S.l.]: GitHub, 2019. https://github.com/qubvel/segmentation_models.pytorch.

[10] COLOMBA, L. et al. A dataset for burned area delineation and severity estimation from satellite imagery. In: Proceedings of the 31st ACM International Conference on Information & Knowledge Management. [S.l.: s.n.], 2022. p. 3893–3897.

[11] AL-DABBAGH, A. M.; ILYAS, M. Uni-temporal Sentinel-2 imagery for wildfire detection using deep learning semantic segmentation models. Geomatics, Natural Hazards and Risk, Taylor & Francis, v. 14, n. 1, p. 2196370, 2023.

[12] RONNEBERGER, O.; FISCHER, P.; BROX, T. U-Net: Convolutional networks for biomedical image segmentation. In: SPRINGER. International Conference on Medical Image Computing and Computer-Assisted Intervention. [S.l.], 2015. p. 234–241.

[13] LIN, T.-Y. et al. Feature pyramid networks for object detection. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). [S.l.: s.n.], 2017. p. 2117–2125.

[14] XIE, E. et al. SegFormer: Simple and efficient design for semantic segmentation with transformers. In: Advances in Neural Information Processing Systems (NeurIPS). [S.l.: s.n.], 2021. v. 34, p. 12077–12090.

[15] HE, K. et al. Deep residual learning for image recognition. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). [S.l.: s.n.], 2016. p. 770–778.

[16] TAN, M.; LE, Q. V. EfficientNet: Rethinking model scaling for convolutional neural networks. In: Proceedings of the 36th International Conference on Machine Learning (ICML). [S.l.]: PMLR, 2019. v. 97, p. 6105–6114. ICML 2019.

[17] FERREIRA, J. et al. Evaluating multiple combinations of models and encoders to segment clouds in satellite images. In: INSTICC. Proceedings of the 19th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 3: VISAPP. [S.l.]: SciTePress, 2024. p. 233–241. ISBN 978-989-758-679-8.