setsci SETSCI Conference Proceedings 2687-5527 SET Technology 10.36287/setsci.21.3.012 36a140d5-ee9a-4ab0-b0cc-586a1ef0bb48 Integrating UAVs and YOLO Deep Learning for Early-Stage Forest Fire Detection Didis Hakan Mert Adibeli Firuze  Boz Ilayda  Azdavay Nesli S.  Mersin University Tarsus University 12 12 2024 21 12 17 © 2024 SET Technology 2024 SET Technology CC BY 4.0

The uncontrolled spread of forest fires, driven by climate change and increasing human activities, poses a significant global challenge, disrupting the delicate balance of ecosystems. These fires, whether caused by natural disasters or human factors, have severe and often irreversible economic, social, and environmental consequences. The rising frequency of forest fires due to both artificial and natural causes is becoming a major international concern. Fire detection methods range from human observation and satellite-based systems to optical smoke detection and watchtowers. While each method offers advantages depending on the application, traditional approaches often suffer from delayed response times, making early intervention difficult. Unmanned aerial vehicles (UAVs) have emerged as a powerful alternative in fire detection, offering the mobility to access difficult terrain and perform rapid, detailed observations over large forested areas. This study explores how forest fires are detected using UAVs integrated with YOLOv4, YOLOv7, and YOLOv9 models, while also examining the limitations of these algorithms. The performance comparison reveals that YOLOv9 outperforms the other models in terms of precision, recall, and speed. With a precision of 0.922, mAP@50 of 0.915, and mAP@50:95 of 0.872, YOLOv9 demonstrates superior performance, exceeding YOLOv7 by 2.78% and YOLOv4 by 9.33%. These findings offer valuable insights into the use of UAVs for fast and accurate fire detection, highlighting their critical role in combating forest fires effectively. 

 forest fire fire detection deep learning unmanned aerial vehicle YOLOv4 YOLOv7 YOLOv9 P. Barmpoutis, P. Papaioannou, K. Dimitropoulos, N. Grammalidis, “A review on early forest fire detection systems using optical remote sensing,” Sensors, vol. 20, no. 6442, 2019. R.S. Priya, K. Vani, "Deep Learning Based Forest Fire Classification and Detection in Satellite Images," 2019 11th International Conference on Advanced Computing (ICoAC), Chennai, India, 2019, pp. 61-65. P. Mittal, R. Singh, A. Sharma, “Deep learning-based object detection in low-altitude UAV datasets: a survey,” Image and Vision Computing, 104, 104046, 2020. M. Bakirci, "A drone-based approach to enhance spatial insight into surrounding air pollutant distributions for healthier indoor environments," Journal of Building Engineering, vol. 87, no. 109023, 2024. https://doi.org/10.1016/j.jobe.2024.109023 M. Bakirci, "Efficient air pollution mapping in extensive regions with fully autonomous unmanned aerial vehicles: A numerical perspective," Science of The Total Environment, vol. 909, no. 168606, 2024. https://doi.org/10.1016/j.scitotenv.2023.168606 M. Bakirci, "Enhancing air pollution mapping with autonomous UAV networks for extended coverage and consistency," Atmospheric Research, vol. 306, no. 107480, 2024. https://doi.org/10.1016/j.atmosres.2024.107480 M. Bakirci, I. Bayraktar, "Harnessing UAV technology and YOLOv9 algorithm for real-time forest fire detection," 2024 International Russian Automation Conference (RusAutoCon), pp. 95-100, Sochi, Russian Federation, 2024. https://doi.org/10.1109/RusAutoCon61949.2024.10694663 M. Bakirci, B. Toptas, "Kinematics and autoregressive model analysis of a differential drive mobile robot," 4th International Congress on Human-Computer Interaction, Optimization and Robotic Applications (HORA), pp. 1-6., Ankara, Turkey, 2022. https://doi.org/10.1109/HORA55278.2022.9800071 Girshick, R. (2015). Fast R-CNN. ArXiv. https://arxiv.org/abs/1504.08083 Ren, S., He, K., Girshick, R., & Sun, J. (2015). Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks. ArXiv. https://arxiv.org/abs/1506.01497 Liu, W., Anguelov, D., Erhan, D., Szegedy, C., Reed, S., Fu, C., & Berg, A. C. (2015). SSD: Single Shot MultiBox Detector. ArXiv. https://doi.org/10.1007/978-3-319-46448-0_2 Jeong, J., Park, H., & Kwak, N. (2017). Enhancement of SSD by concatenating feature maps for object detection. ArXiv. https://arxiv.org/abs/1705.09587 Duan, K., Bai, S., Xie, L., Qi, H., Huang, Q., & Tian, Q. (2019). CenterNet: Keypoint Triplets for Object Detection. ArXiv. https://arxiv.org/abs/1904.08189 Redmon, J., Divvala, S., Girshick, R., & Farhadi, A. (2015). You Only Look Once: Unified, Real-Time Object Detection. ArXiv. https://arxiv.org/abs/1506.02640 P. Soviany and R. T. Ionescu, "Optimizing the Trade-Off between Single-Stage and Two-Stage Deep Object Detectors using Image Difficulty Prediction," 2018 20th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing (SYNASC), Timisoara, Romania, 2018, pp. 209-214. T. Diwan, G. Anirudh, J.V. Tembhurne, "Object detection using YOLO: challenges, architectural successors, datasets and applications," Multimed Tools Appl, vol. 82, pp. 9243–9275, 2023. https://doi.org/10.1007/s11042-022-13644-y Bochkovskiy, A., Wang, C., & Liao, H. (2020). YOLOv4: Optimal Speed and Accuracy of Object Detection. ArXiv. https://arxiv.org/abs/2004.10934 R. Niu, Y. Qu and Z. Wang, "UAV Detection Based on Improved YOLOv4 Object Detection Model," 2021 2nd International Conference on Big Data & Artificial Intelligence & Software Engineering (ICBASE), Zhuhai, China, 2021, pp. 25-29. Wang, C., Yeh, I., & Liao, H. (2024). YOLOv9: Learning What You Want to Learn Using Programmable Gradient Information. ArXiv. https://arxiv.org/abs/2402.13616 M. Bakirci, I. Bayraktar, "Transforming aircraft detection through LEO satellite imagery and YOLOv9 for improved aviation safety," 2024 26th International Conference on Digital Signal Processing and its Applications (DSPA), pp. 1-6, Moscow, Russian Federation, 2024. ttps://doi.org/10.1109/DSPA60853.2024.10510106