Wildfire detection using hybrid vision transformer and MobileNetV2 with XAI

Main Article Content

Saif Mostasir Rohan
https://orcid.org/0009-0007-1756-8400
Md. Sayem Kabir
Tasnim Sultana Sintheia
Kazi Tanvir
Dipta Gomes
Mahfujur Rahman
https://orcid.org/0000-0002-3245-8767
Akinul Islam Jony

Abstract

Wildfires cause severe ecological and economic damage, and limiting it requires rapid and accurate detection. This study presents an image classification framework for wildfire detection in RGB images from the FLAME 3 dataset, captured by an unmanned aerial vehicle. The preprocessing pipeline applies gamma correction, sharpening, contrast-limited adaptive histogram equalisation (CLAHE) and class-balanced data augmentation. The proposed model combines a vision transformer (ViT) and MobileNetV2 in parallel branches, fusing global contextual features with local spatial features. Explainable artificial intelligence techniques (Grad-CAM, Grad-CAM++ and LIME) are used to visualise the image regions that influence the predictions. The model is compared with nine convolutional and five transformer-based baselines and evaluated with accuracy, Cohen’s κ, the Matthews correlation coefficient (MCC), the critical success index (CSI) and related metrics. Trained with the AdamW optimiser, the ViT–MobileNetV2 model achieves a test accuracy of 98.44%, a macro-average F1-score of 0.984 and a recall of 1.000 for the fire class on the FLAME 3 RGB classification subset. Inference takes 10–15 ms per image on an NVIDIA T4 GPU in half precision and 50–70 ms on an x86 CPU with INT8 quantisation. The results apply to the evaluated dataset and experimental protocol.

Views: 0Downloads: 0

Downloads

Download data is not yet available.

Article Details

Section

Articles

How to Cite

Rohan, S. M., Kabir, M. S., Sintheia, T. S., Tanvir, K., Gomes, D., Rahman, M., & Jony, A. I. (2026). Wildfire detection using hybrid vision transformer and MobileNetV2 with XAI. Journal of Edge Computing. https://doi.org/10.55056/jec.1157

References

Baker, S.J., Perry, M.C., Betts, R.A., Schoenecker, J. and Pellegrini, A.F.A., 2025. Spikes in UK wildfire emissions driven by peatland fires in dry years. Environmental Research Letters, 20(3), 034028. Available from: https://doi.org/10.1088/1748-9326/adafc6.

Chattopadhay, A., Sarkar, A., Howlader, P. and Balasubramanian, V.N., 2018. Grad-CAM++: Generalized Gradient-Based Visual Explanations for Deep Convolutional Networks. 2018 IEEE Winter Conference on Applications of Computer Vision (WACV). pp.839–847. Available from: https://doi.org/10.1109/WACV.2018.00097.

Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., Dehghani, M., Minderer, M., Heigold, G., Gelly, S., Uszkoreit, J. and Houlsby, N., 2021. An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale. 9th International Conference on Learning Representations, ICLR 2021. OpenReview.net. Available from: https://openreview.net/forum?id=YicbFdNTTy.

Dozat, T., 2016. Incorporating Nesterov Momentum into Adam. International Conference on Learning Representations (ICLR) Workshop Track. Available from: https://openreview.net/pdf/OM0jvwB8jIp57ZJjtNEZ.pdf.

Ghali, R., Akhloufi, M.A., Jmal, M., Souidene Mseddi, W. and Attia, R., 2021. Wildfire Segmentation Using Deep Vision Transformers. Remote Sensing, 13(17), 3527. Available from: https://doi.org/10.3390/rs13173527.

Ghali, R., Akhloufi, M.A. and Mseddi, W.S., 2022. Deep Learning and Transformer Approaches for UAV-Based Wildfire Detection and Segmentation. Sensors, 22(5), 1977. Available from: https://doi.org/10.3390/s22051977.

Glorot, X., Bordes, A. and Bengio, Y., 2011. Deep Sparse Rectifier Neural Networks. In: G. Gordon, D. Dunson and M. Dudík, eds. Proceedings of the Fourteenth International Conference on Artificial Intelligence and Statistics, Proceedings of machine learning research, vol. 15. Fort Lauderdale, FL, USA: PMLR, pp.315–323. Available from: https://proceedings.mlr.press/v15/glorot11a.html.

Guan, Z., Miao, X., Mu, Y., Sun, Q., Ye, Q. and Gao, D., 2022. Forest Fire Segmentation from Aerial Imagery Data Using an Improved Instance Segmentation Model. Remote Sensing, 14(13), 3159. Available from: https://doi.org/10.3390/rs14133159.

Hopkins, B., O’Neill, L., Marinaccio, M., Afghah, F., Rowell, E., Parsons, R. and Flanary, S., 2024. FLAME 3 Computer Vision Subset (Sycan Marsh). Available from: https://www.kaggle.com/dsv/8480870.

Ioffe, S. and Szegedy, C., 2015. Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift. In: F. Bach and D. Blei, eds. Proceedings of the 32nd International Conference on Machine Learning, Proceedings of machine learning research, vol. 37. Lille, France: PMLR, pp.448–456. Available from: https://proceedings.mlr.press/v37/ioffe15.html.

Islam, A.M., Masud, F.B., Ahmed, M.R., Jafar, A.I., Ullah, J.R., Islam, S., Shatabda, S. and Islam, A.K.M.M., 2023. An Attention-Guided Deep-Learning-Based Network with Bayesian Optimization for Forest Fire Classification and Localization. Forests, 14(10), 2080. Available from: https://doi.org/10.3390/f14102080.

Khan, A., Hassan, B., Khan, S., Ahmed, R. and Abuassba, A., 2022. DeepFire: A Novel Dataset and Deep Transfer Learning Benchmark for Forest Fire Detection. Mobile Information Systems, 2022, 5358359. Available from: https://doi.org/10.1155/2022/5358359.

Kingma, D.P. and Ba, J., 2015. Adam: A Method for Stochastic Optimization. In: Y. Bengio and Y. LeCun, eds. 3rd International Conference on Learning Representations, ICLR 2015, San Diego, CA, USA, May 7–9, 2015, Conference Track Proceedings. Available from: https://doi.org/10.48550/arXiv.1412.6980.

Knutzen, F., Averbeck, P., Barrasso, C., Bouwer, L.M., Gardiner, B., Grünzweig, J.M., Hänel, S., Haustein, K., Johannessen, M.R., Kollet, S., Müller, M.M., Pietikäinen, J.P., Pietras-Couffignal, K., Pinto, J.G., Rechid, D., Rousi, E., Russo, A., Suarez-Gutierrez, L., Veit, S., Wendler, J., Xoplaki, E. and Gliksman, D., 2025. Impacts on and damage to European forests from the 2018–2022 heat and drought events. Natural Hazards and Earth System Sciences, 25(1), pp.77–117. Available from: https://doi.org/10.5194/nhess-25-77-2025.

Laghari, A.A., Jumani, A.K., Laghari, R.A. and Nawaz, H., 2023. Unmanned aerial vehicles: A review. Cognitive Robotics, 3, pp.8–22. Available from: https://doi.org/10.1016/j.cogr.2022.12.004.

Li, P., He, X., Cheng, X., Qiao, M., Song, D., Chen, M., Zhou, T., Li, J., Guo, X., Hu, S. and Tian, Z., 2022. An improved categorical cross entropy for remote sensing image classification based on noisy labels. Expert Systems with Applications, 205, 117296. Available from: https://doi.org/10.1016/j.eswa.2022.117296.

Loshchilov, I. and Hutter, F., 2019. Decoupled Weight Decay Regularization. 7th International Conference on Learning Representations, ICLR 2019, New Orleans, LA, USA, May 6–9, 2019. OpenReview.net. Available from: https://openreview.net/forum?id=Bkg6RiCqY7.

Pereira, G.A. and Hussain, M., 2024. A Review of Transformer-Based Models for Computer Vision Tasks: Capturing Global Context and Spatial Relationships. arXiv preprint arXiv:2408.15178. Available from: https://doi.org/10.48550/arXiv.2408.15178.

Prakash, M., Neelakandan, S., Tamilselvi, M., Velmurugan, S., Baghavathi Priya, S. and Ofori Martinson, E., 2023. Deep Learning-Based Wildfire Image Detection and Classification Systems for Controlling Biomass. International Journal of Intelligent Systems, 2023(1), 7939516. Available from: https://doi.org/10.1155/2023/7939516.

Reza, A.M., 2004. Realization of the Contrast Limited Adaptive Histogram Equalization (CLAHE) for Real-Time Image Enhancement. Journal of VLSI Signal Processing Systems for Signal, Image and Video Technology, 38, pp.35–44. Available from: https://doi.org/10.1023/B:VLSI.0000028532.53893.82.

Ribeiro, M.T., Singh, S. and Guestrin, C., 2016. “Why Should I Trust You?”: Explaining the Predictions of Any Classifier. Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. pp.1135–1144. Available from: https://doi.org/10.1145/2939672.2939778.

Sandler, M., Howard, A., Zhu, M., Zhmoginov, A. and Chen, L.C., 2018. MobileNetV2: Inverted Residuals and Linear Bottlenecks. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp.4510–4520. Available from: https://doi.org/10.1109/CVPR.2018.00474.

Selvaraju, R.R., Cogswell, M., Das, A., Vedantam, R., Parikh, D. and Batra, D., 2017. Grad-CAM: Visual Explanations from Deep Networks via Gradient-Based Localization. 2017 IEEE International Conference on Computer Vision (ICCV). pp.618–626. Available from: https://doi.org/10.1109/ICCV.2017.74.

Tang, Y., Feng, H., Chen, J. and Chen, Y., 2021. ForestResNet: A Deep Learning Algorithm for Forest Image Classification. Journal of Physics: Conference Series, 2024(1), 012053. Available from: https://doi.org/10.1088/1742-6596/2024/1/012053.

Tanvir, K., Kabir, M.S., Anik, Z.H., Hasan, M.R., Tushar, M.T. and Rahman, R.B., 2024. Forest Fire Detection Using Ensemble Deep Learning Model With XAi. 2024 IEEE International Conference on Computing, Applications and Systems (COMPAS). pp.1–6. Available from: https://doi.org/10.1109/COMPAS60761.2024.10796582.

United Nations Office for Disaster Risk Reduction, 2023. Regional Assessment Report on Disaster Risk Reduction 2023: Europe and Central Asia. United Nations Office for Disaster Risk Reduction. Available from: https://www.undrr.org/report/rar-2023-europe-and-central-asia.

Yang, S., Huang, Q. and Yu, M., 2024. Advancements in remote sensing for active fire detection: A review of datasets and methods. Science of the Total Environment, 943, 173273. Available from: https://doi.org/10.1016/j.scitotenv.2024.173273.

Zhang, L., Wang, M., Fu, Y. and Ding, Y., 2022. A Forest Fire Recognition Method Using UAV Images Based on Transfer Learning. Forests, 13(7), 975. Available from: https://doi.org/10.3390/f13070975.