Journal of the Korean Association of Geographic Information Studies (한국지리정보학회지)

The Korean Association of Geographic Information Studies (한국지리정보학회)
권호 표지
DOI QR코드

DOI QR Code

A study on the development of an automatic detection algorithm for trees suspected of being damaged by forest pests

산림병해충 피해의심목 자동탐지 알고리즘 개발 연구

  • Received : 2022.11.02
  • Accepted : 2022.12.09
  • Published : 2022.12.31
Download PDF
⟨ Previous Next ⟩

Abstract

Recently, the forests in Korea have accumulated damage due to continuous forest disasters, and the need for technologies to monitor forest managements is being issued. The size of the affected area is large terrain, technologies using drones, artificial intelligence, and big data are being studied. In this study, a standard dataset were conducted to develop an algorithm that automatically detects suspicious trees damaged by forest pests using deep learning and drones. Experiments using the YOLO model among object detection algorithm models, the YOLOv4-P7 model showed the highest recall rate of 69.69% and precision of 69.15%. It was confirmed that YOLOv4-P7 should be used as an automatic detection algorithm model for trees suspected of being damaged by forest pests, considering the detection target is an ortho-image with a large image size.

최근 우리나라의 산림은 지속적인 산림재해로 인해 피해가 누적되고 있어 산림을 관리하기 위한 모니터링 기술이 조명받고 있으며, 산림재해 피해대상지의 규모가 큰 지형 특성으로 인해 드론, 인공지능, 빅데이터 등을 활용한 기술들이 연구되고 있다. 본 연구에서는 산림재해의 병해충을 모니터링하기 위해 딥러닝과 드론을 활용하여 산림 병해충 피해 의심목을 자동으로 탐지하는 산림 병해충 자동탐지 알고리즘 개발을 위한 표준 데이터 세트를 구축하였다. 객체검출 알고리즘으로서 YOLO 알고리즘을 활용한 실험결과에서는 YOLOv4-P7 모델이 재현율 69.69%와 정밀도 69.15%로 가장 높게 나타났으며, 이미지 사이즈가 큰 정사영상인 검출대상임을 고려할 때 산림병해충 피해의심목 자동탐지 알고리즘으로 YOLOv4-P7이 적합함을 확인하였다.

Keywords

References

  1. Francois Chollet. 2018. Deep learning from the creator of Keras, gilbut
  2. Jang, K.M. 2021. A Study on the Deep Learning -based Tree Species Classification by using High-resolution Orthophoto Images. 한국지리정보학회지, 3(24):1-9.
  3. Johan Marbinah. 2021. Hybrid pool based deep activelearning for object detection using intermediate network embeddings, KTH ROYAL INSTITUTE OF TECHNOLOGY.
  4. Kim, J.H., Lee, T.H. Yamin Han, and Byun, H.J. 2021. A Study on the Design and Implementation of Multi-Disaster Drone System Using Deep Learning-Based Object Recognition and Optimal Path Planning. 정보처리학회논문지. 컴퓨터 및 통신시스템, 10(4):117-122.
  5. Lee, J.H., Kim, J.C. and Seo. D.H. 2017. A study on image caption algorithm based on object detection. Journal of Advanced Marine Engineering and Technology, 41(7): 683-689.
  6. Lee, Y.J. 2020. Surveying and Geoinformatics, Gyomoon co.
  7. Lee. H.D., 2022. Victim Tree Detection System of Drone Images Based on 3D GIS and Deep Leaming Method. Department of Construction Engineering Graduate School, Kyungil University.
  8. Lim, E.T. and Do, M.S. 2021. Pine Wilt Disease Detection Based on Deep Learning Using an Unmanned Aerial Vehicle. JOURNAL OF THE KOREAN SOCIETY OF CIVIL ENGINEERS, 41(3):317-325.
  9. Park, J.I., and Lee, M.S. 2021. Analysis of trees damaged by pine wilt nematodes using unmanned aerial images. 한국지적정보학회 학술발표대회 논문집, 13-16.
  10. Redmon, J., Divvala, S., Girshick, R., & Farhadi, A. 2016. You only look once: Unified, real-time object detection. In Proceedings of the IEEE conference on computer vision and pattern recognition pp.779-788.
  11. SON, B., LEE, Y., & IM, J. (2021). Classification of Urban Green Space Using Airborne LiDAR and RGB Ortho Imagery Based on Deep Learning. Journal of the Korean Association of Geographic Information Studies, 24(3), 83-98.