한국군사과학기술학회지 (Journal of the Korea Institute of Military Science and Technology)

  • 제28권3호
  • /
  • Pages.254-263
  • /
  • 2025
  • /
  • 1598-9127(pISSN)
  • /
  • 2636-0640(eISSN)
한국군사과학기술학회 (The Korea Institute of Military Science and Technology)
권호 표지
DOI QR코드

DOI QR Code

산불 대응을 위한 딥러닝 활용 스마트 공대지 유도 시스템

Smart Air-to-Ground Guidance System using Deep Learning for Wildfire Response

  • 박경필 (연세대학교 전기전자공학과) ;
  • 김동빈 (하트포트대학교 전기컴퓨터공학과) ;
  • 남태식 (연세대학교 전기전자공학과) ;
  • 육종관 (연세대학교 전기전자공학과)
  • Gyeongphil Park (Department of Electrical and Electronic Engineering, Yonsei University) ;
  • Dongbin Kim (Department of Electrical and Computer Engineering,University of Hartford) ;
  • Taesik Nam (Department of Electrical and Electronic Engineering, Yonsei University) ;
  • Jonggwan Yook (Department of Electrical and Electronic Engineering, Yonsei University)
  • 투고 : 2025.01.24
  • 심사 : 2025.03.26
  • 발행 : 2025.06.05
PDF 다운로드
⟨ 이전 논문 다음 논문 ⟩

초록

Wildfires pose significant environmental and economic challenges, intensifying due to climate change. This study introduces the Guided Fire Extinguishing Device(GFED), an autonomous air-to-ground system leveraging deep learning for precise wildfire detection, tracking, and response. Using a deep learning model for object detection and the Nona Filter, a priority-based target tracking algorithm, GFED achieves robust performance in adverse conditions such as strong winds, night view, and heavy smoke. Key contributions include optimized deep learning model training, precise mechanical trajectory control, and real-time tracking capabilities. The aerodynamic design, optimized for payload capacity and stability, ensures scalability and reliability. Experimental results demonstrate the effectiveness of GFED, achieving a mean average precision(mAP) of 94.5 % for fire detection. This transformative approach enhances global wildfire response efforts, offering improved safety and efficiency in combating the growing threat of wildfires.

키워드

과제정보

이 논문은 정부(행정안전부, 과학기술정보통신부, 산업통상자원부, 소방청)의 재원으로 한국산업기술기획평가원의 지원을 받아 수행된 연구임(과제명: 난접근성 화재 대응을 위한 가스하이드레이트 소화탄 및 화재 진압 기술개발사업, 과제번호 : 2760000002).

참고문헌

  1. A. Sullivan et al., "Spreading like wildfire: The rising threat of extraordinary landscape fires," Tech. Rep., UNEP, 2022.
  2. J. Redmon, S. Divvala, R. Girshick, and A. Farhadi, "You only look once: Unified, real-time object detection," in Proc. IEEE Conf. Comput. Vis. Pattern Recognit.(CVPR), pp. 779-788, 2016.
  3. I. Jeong, "A study on carbon monoxide and other gases during the fire test of A-class and B-class(1st rating)," Fire Sci. Eng., Vol. 27, No. 5, pp. 44-56, 2013. https://doi.org/10.7731/KIFSE.2013.27.5.44
  4. J. Terven and D. Córdova-Esparza, "A comprehensive review of YOLO architectures in computer vision: From YOLOv1 to YOLOv8 and YOLO-NAS," Mach. Learn. Knowl. Extr., Vol. 5, No. 4, pp. 1680-1716, 2023. https://doi.org/10.3390/make5040083
  5. MMYOLO Contributors, "YOLOv8 by MMYOLO," GitHub repository, [Online]. Available: https://github.com/open-mmlab/mmyolo/tree/main/configs/yolov8. Accessed: Mar. 2, 2025.
  6. V. E. Sathishkumar, J. Cho, M. Subramanian, and O. S. Naren, "Forest fire and smoke detection using deep learning-based learning without forgetting," Fire ecology, Vol. 19, No. 1, p. 9, 2023. https://doi.org/10.1186/s42408-022-00165-0
  7. D. Han, J. Kim, and J. Kim, "Deep pyramidal residual networks," in 2017 IEEE Conference on Computer Vision and Pattern Recognition(CVPR), pp. 6307-6315, 2017.
  8. D. S. Bolme, J. R. Beveridge, B. A. Draper, and Y. M. Lui, "Visual object tracking using adaptive correlation filters," in Proc. IEEE Conf. Comput. Vis. Pattern Recognit.(CVPR), pp. 2544-2550, 2010.
  9. Z. Kalal, K. Mikolajczyk, and J. Matas, "Trackinglearning-detection," IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 34, No. 7, pp. 1409-1422, 2012. https://doi.org/10.1109/TPAMI.2011.239
  10. G. Park and D. Kim, "Design of artificial intelligence air-to-ground guidance control device for wildfire extinguishing," in Proc. Korean Inst. Mil. Sci. Technol.(KIMST) Conf., pp. 1193-1194, 2024.
  11. M. S. Kim and H. B. Kim, Aerodynamics, 2nd ed., Seoul, Korea: SeHwa Publishing, 2012.
  12. A. K. Wong and N.-K. Fong, "Study of pool fire heat release rate using video fire detection," in Proc. Int. Conf. Fire Safety, 2014.
  13. S. M. Jo and O. J. Kwon, "Numerical study about behavior of an ejecting projectile for varying initial conditions," J. Korean Soc. Aeronaut. Space Sci., Vol. 47, No. 11, pp. 761-767, 2019.  https://doi.org/10.5139/JKSAS.2019.47.11.761