DOI QR코드

DOI QR Code

Study on the Application of RT-DETR to Monitoring of Coastal Debris on Unmanaged Coasts

비관리 해변의 해안 쓰레기 모니터링을 위한 RT-DETR 적용 방안 연구

  • Ye-Been Do ;
  • Hong-Joo Yoon (Divsion of Earth Environmental System Science Major of Geomatics Engineering, Pukyong National University)
  • 도예빈 (부경대학교 지구환경시스템과학부 공간정보시스템공학전공) ;
  • 윤홍주 (부경대학교 지구환경시스템과학부 위성정보융합공학전공)
  • Received : 2024.02.29
  • Accepted : 2024.04.12
  • Published : 2024.04.30

Abstract

To improve the monitoring of Coastal Debris in the South Korea, which is difficult to estimate due to limited resources and vertex-based surveys, an approach based on UAV(Unmanned Aerial Vehicle) images and the RT-DETR(Realtime DEtection TRansformer) model was proposed for detecting Coastal Debris. By comparing to field investigation, the study suggested the possibility of quantitatively detecting coastal garbage and estimating the total capacity of garbage deposited on the natural coastline of the South Korea. The RT-DETR model achieved an accuracy of 0.894 for mAP@0.5 and 0.693 for mAP@0.5:0.95 in training. When applied to unmanaged coasts, the accuracy for the total number of coastal debris items was 72.9%. It is anticipated that if guidelines for defining monitoring of unmanaged coasts are established alongside this research, it should be possible to estimate the total capacity of the deposited coastal debris in the South Korea.

한정된 정점과 인력 기반의 조사로 실제 국내 표착되는 해안 쓰레기의 총용량 추정이 어려운 우리나라의 해안 쓰레기 모니터링 방식 개선을 위해 비관리 해변에서 UAV(: Unmanned Aerial Vehicle) 이미지와 RT-DETR 모델을 기반으로 해안 쓰레기 탐지하고 현장 조사와의 비교 연구로 해안 쓰레기의 정량적 탐지 및 자연 해안선 기준 우리나라에 표착되는 전체 쓰레기 총용량 추정 가능성을 제시하였다. RT-DETR(: Realtime DEtection TRansformer) 모델 학습 결과 mAP@0.5는 0.894, mAP@0.5:0.95는 0.693의 정확도를 보였다. 모델을 비관리 해변에 적용한 전체 해안 쓰레기 개수에 대한 정확도는 72.9%로 나타났다. 본 연구와 비관리 해변에 대한 모니터링을 정의하는 관리지침 마련 연구가 동반된다면 우리나라에 표착되는 전체 해안 쓰레기의 총 용량 추정이 가능할 것으로 기대된다.

Keywords

Acknowledgement

위 논문은 부경대학교 자율창의학술연구비 (2023년)에 의하여 연구되었음

References

  1. Ministry of Oceans and Fisheries, 2021, 2020 National coastal garbage control and monitoring investigation service, Final report, Korea Marine Environment Management Corporation, Korea.
  2. S. Jang, J. Park, Y. Chung, D. Kim and H. Yoon, "A Study on the Inflow and Seasonal Characteristics of Foreign Marine Debris in the Coastal Area of the West Sea," J. the Korean Society for Marine Environment & Energy, vol 15, no. 2, May. 2012, pp. 89-100.
  3. F. Thevenon, C. Carroll, J. Sousa, "Plastic debris in the ocean: the characterization of marine plastics and their environmental impacts, situation analysis report," Gland, Switzerland: IUCN 52, 2014. pp. 549-562.
  4. J. Park, D. Kim, H. Yoon, and W. Seo, "A Study on Identification of Characteristics of Spatial Distribution for Submerged Marine Debris," J. of the Korea Institute of Electronic Communication Sciences(KIECS), vol. 11, no. 5, 2016, pp. 539-544. https://doi.org/10.13067/JKIECS.2016.11.5.539
  5. K. Choi, "A Coastal Garbage Monitoring System Using Drones and AI Technologies: Focusing on the Case of Jeju Province," J. Korean Society for Geospatial Information Science, vol 29, no. 4, Dec. 2021, pp. 127-138. https://doi.org/10.7319/kogsis.2021.29.4.127
  6. Tekman, M. B., Walther, B. A., Peter, C., Gutow, L. and Bergmann, M. (2022): Impacts of plastic pollution in the oceans on marine species, biodiversity and ecosystems, 1-221, WWF Germany, Berlin. Doi: 10.5281/zenodo.5898684.
  7. R. Pfeiffer, G. Valentino, R. A. Farrugia, E. Colica, S. D'Amico, and S. Calleja, "Detecting beach litter in drone images using deep learning." In 2022 IEEE IInt. Workshop on Metrology for the Sea; Learning to Measure Sea Health Parameters (MetroSea), Oct. 2022, pp. 28-32.
  8. V. M. Scarrica, P. P. Aucelli, C. Cagnazzo, A. Casolaro, P. Fiore, M. La Salandra, A. Rizzo, g. Scardino, G. Scicchitano and A. Staiano, "A novel beach litter analysis system based on UAV images and Convolutional Neural Networks," Ecological Informatics, vol. 72, Dec. 2022, 101875.
  9. N. Maharjan, H. Miyazaki, B. M. Pati, M. N. Dailey, S. Shrestha, and T. Nakamura, "Detection of river plastic using UAV sensor data and deep learning," Remote Sensing, vol. 14, no. 13, June 2022, 3049.
  10. B. Kim, M. Park, J. Kim, Y. Do, S. Oh and H. Yoon, "Analysis Temporal Variations Marine Debris by using Raspberry Pi and YOLOv5," J. of the Korea Institute of Electronic Communication Sciences(KIECS) vol. 17, no. 6, Dec. 2021, pp. 1249-1258.
  11. Y. Lee and D. Jeong, "On Drone Altitude and Trash Recognition Rat," J. of KIIT, vol. 19, no. 1, Jan. 2021, pp. 33-42. https://doi.org/10.14801/jkiit.2021.19.1.33
  12. Y. Zhao, W. Lv, S. Xu, J. Wei, G. Wang, Q. Dang, Y. Liu and J. Chen, "DETRs Beat YOLOs on Real-time Object Detection," Apr. 2024., arXiv preprint arXiv:2304.08069.
  13. H. Myung and J. Song, "Deep Learning-based Poultry Object Detection Algorithm," J. of Digital Contents Society, Vol 23, no. 7, July. 2022, pp.1323-1330. https://doi.org/10.9728/dcs.2022.23.7.1323