Fire Detection from Forest Images Using Multiple Deep Learning Models
Abstract views: 11
/ 
PDF downloads: 10
Keywords:
Artificial Intelligence, Deep Learning, Fire Detection, Image ClassificationAbstract
Forest fires can occur from natural or unnatural causes. While unnatural fires usually consist of
flammable materials thrown around unconsciously; Natural fires occur regularly depending on forest
cover, forest type, soil type and climate. Within the scope of this study, fire detection processes were
carried out with artificial intelligence using open source dataset in order to detect fires in forests. The
dataset used includes fire and non-fire images. There are 480 images in total, 240 for each class in the
binary classification study. For classification, special attention has been paid to the equal distribution of
classes so that the network training in deep learning models can be carried out in the best way. The
dataset is randomly split into 80% train, 20% validation. The dataset distribution consists of 384 images
for training and 96 images for validation. Artificial intelligence-based deep learning models used for
classification are Residual Networks (ResNet), Bidirectional Encoder representation from Image
Transformers (BEiT) and Swin Transformer. When the fire detection results were examined, it was
observed that the classification accuracies were above 87% and the f1-score was above 86%. In future
studies, hybrid and/or ensemble models can be developed by using more deep learning models to further
improve the detection process.
Downloads
References
Johnson, E. A., & Miyanishi, K. (2001). Forest fires: Behavior and ecological effects. Academic Press.
Abdusalomov, A. B., Islam, B. M. S., Nasimov, R., Mukhiddinov, M., & Whangbo, T. K. (2023). An improved forest fire detection method based on the detectron2 model and a deep learning approach. Sensors, 23(3), 1512.
Yar, H., Khan, Z. A., Rida, I., Ullah, W., Kim, M. J., & Baik, S. W. (2024). An efficient deep learning architecture for effective fire detection in smart surveillance. Image and Vision Computing, 145, 104989.
Shamta, I., & Demir, B. E. (2024). Development of a deep learning-based surveillance system for forest fire detection and monitoring using UAV. Plos one, 19(3), e0299058.
Titu, M. F. S., Pavel, M. A., Michael, G. K. O., Babar, H., Aman, U., & Khan, R. (2024). Real-Time Fire Detection: Integrating Lightweight Deep Learning Models on Drones with Edge Computing. Drones, 8(9), 483.
Jin, L., Yu, Y., Zhou, J., Bai, D., Lin, H., & Zhou, H. (2024). SWVR: a lightweight deep learning algorithm for forest fire detection and recognition. Forests, 15(1), 204.
Zhao, H., Jin, J., Liu, Y., Guo, Y., & Shen, Y. (2024). FSDF: A high-performance fire detection framework. Expert Systems with Applications, 238, 121665.
Cheknane, M., Bendouma, T., & Boudouh, S. S. (2024). Advancing fire detection: Two-stage deep learning with hybrid feature extraction using faster R-CNN approach. Signal, Image and Video Processing, 18(6), 5503-5510.
El-Madafri, I., Peña, M., & Olmedo-Torre, N. (2024). Dual-Dataset Deep Learning for Improved Forest Fire Detection: A Novel Hierarchical Domain-Adaptive Learning Approach. Mathematics, 12(4), 534.
Yunusov, N., Islam, B. M. S., Abdusalomov, A., & Kim, W. (2024). Robust Forest Fire Detection Method for Surveillance Systems Based on You Only Look Once Version 8 and Transfer Learning Approaches. Processes, 12(5), 1039.
Sharma, A., Kumar, R., Kansal, I., Popli, R., Khullar, V., Verma, J., & Kumar, S. (2024). Fire detection in urban areas using multimodal data and federated learning. Fire, 7(4), 104.
Yang, W., Wu, Y., & Chow, S. K. K. (2024). Deep Learning Method for Real-Time Fire Detection System for Urban Fire Monitoring and Control. International Journal of Computational Intelligence Systems, 17(1), 216.
Yu, R., & Kim, K. (2024). A Study of Novel Initial Fire Detection Algorithm Based on Deep Learning Method. Journal of Electrical Engineering & Technology, 19(6), 3675-3686.
Yar, H., Khan, Z. A., Hussain, T., & Baik, S. W. (2024). A modified vision transformer architecture with scratch learning capabilities for effective fire detection. Expert Systems with Applications, 252, 123935.
Buriboev, A. S., Rakhmanov, K., Soqiyev, T., & Choi, A. J. (2024). Improving Fire Detection Accuracy through Enhanced Convolutional Neural Networks and Contour Techniques. Sensors, 24(16), 5184.
Xu, Y., Li, J., Zhang, L., Liu, H., & Zhang, F. (2024). CNTCB-YOLOv7: An effective forest fire detection model based on ConvNeXtV2 and CBAM. Fire, 7(2), 54.
Kaggle. (n.d.). Fire dataset. Retrieved February 1, 2025, from https://www.kaggle.com/datasets/phylake1337/fire-dataset
He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 770-778).
Bao, H., Dong, L., Piao, S., & Wei, F. (2021). Beit: Bert pre-training of image transformers. arXiv preprint arXiv:2106.08254.
Liu, Z., Lin, Y., Cao, Y., Hu, H., Wei, Y., Zhang, Z., ... & Guo, B. (2021). Swin transformer: Hierarchical vision transformer using shifted windows. In Proceedings of the IEEE/CVF international conference on computer vision (pp. 10012-10022).
Hugging Face. (n.d.). Hugging Face models. Retrieved February 15, 2025, from https://huggingface.co/models
