DOI QR코드

DOI QR Code

A Study on the Implementation of Real-Time Marine Deposited Waste Detection AI System and Performance Improvement Method by Data Screening and Class Segmentation

데이터 선별 및 클래스 세분화를 적용한 실시간 해양 침적 쓰레기 감지 AI 시스템 구현과 성능 개선 방법 연구

  • 왕태수 (동의대학교 컴퓨터 공학과) ;
  • 오세영 (동의대학교 IT융합학과) ;
  • 이현서 (동의대학교 산업ICT기술공학과) ;
  • 최동규 (동의대학교 스마트IT연구소) ;
  • 장종욱 (동의대학교 컴퓨터 공학과) ;
  • 김민영 (동의대학교 ICT융복합연구소)
  • Received : 2022.04.28
  • Accepted : 2022.05.08
  • Published : 2022.05.31

Abstract

Marine deposited waste is a major cause of problems such as a lot of damage and an increase in the estimated amount of garbage due to abandoned fishing grounds caused by ghost fishing. In this paper, we implement a real-time marine deposited waste detection artificial intelligence system to understand the actual conditions of waste fishing gear usage, distribution, loss, and recovery, and study methods for performance improvement. The system was implemented using the yolov5 model, which is an excellent performance model for real-time object detection, and the 'data screening process' and 'class segmentation' method of learning data were applied as performance improvement methods. In conclusion, the object detection results of datasets that do screen unnecessary data or do not subdivide similar items according to characteristics and uses are better than the object recognition results of unscreened datasets and datasets in which classes are subdivided.

해양침적쓰레기는 유령어업으로 인한 폐어구들로 인해 많은 피해와 쓰레기 추정량 편차 증가 등의 문제를 일으키는 주요 원인이 된다. 본 논문에서는 폐어구 사용량, 유통량, 유실량, 회수량에 대한 실태 파악을 위해 실시간 해양침적쓰레기 감지 인공지능 시스템을 구현하고, 성능 개선을 위한 방법에 대해 연구한다. 실시간 객체인식에 우수한 성능모델인 yolov5모델을 활용하여 시스템을 구현하였고, 성능개선 방법으로는 학습데이터의 '데이터 선별 과정'과 '클래스 세분화' 방법을 적용하였다. 결론적으로 비선별된 데이터셋과 클래스가 세분화된 데이터셋의 객체인식 결과보다 불필요한 데이터를 선별하거나 특징 및 용도에 따라 유사 항목을 세분화 하지 않은 데이터셋의 객체인식 결과는 해양침적쓰레기 인식에 개선된 결과를 보인다.

Keywords

Acknowledgement

본 논문(저서)는 부산광역시 및 (재)부산인재평생교육진흥원의 BB21플러스 사업으로 지원된 연구임. 또한 본 연구는 과학기술정보통신부 및 정보통신기획평가원의 지역지능화 혁신인재양성(Grand ICT연구센터) 사업의 연구결과로 수행되었음.(IITP-2022-2016-0-00318)

References

  1. Y. J. Kim and S. M. Jeong, "Consideration on the Standing Stock of Marine Wastes" in Journal of Korea Society of Waste Management, Vol. 33, No. 6, pp. 598-605, Sep. 2016. (DOI : 10.9786/kswm.2016.33.6.598)
  2. Marine Environment Information Portal, "The justice of marine waste." Internet: https://www.meis.go.kr/mli/intro/learn.do,[Jan.5,2022]
  3. KOEM, "The final report of 2020 National Constal Waste Monitoring." Internet: https://www.koem.or.kr/site/koem/ex/board/View.do?cbIdx=370&bcIdx=30182, Apr.20,2021[Jan.5,2022]
  4. Ministry of Oceans and Fisheries "The 1st Basic Plan for the Management of Marine Waste and Marine Pollution Deposits (2021-2030)." Internet: https://www.mof.go.kr/article/view.do?articleKey=39282&boardKey=22&menuKey=851¤tPageNo=1, Jun.30,2021[Jan.5,2022]
  5. S. H. Kim, Y. S. Yang, J. H. Jeong, M. S. Song, D. H. Lee, M. B. Choi, S. H. Cha, C. H. Lee and H. S. Kim, "Verification on the Ghost Fishing by Derelict Fishing Gears through the Tank Experiments" in Journal of Fishries and Marine Sciences Education, Vol. 28, No. 5, pp. 1258-1265, Oct. 2016. (DOI : https://doi.org/10.13000/JFMSE.2016.28.5.1258)
  6. S. D. Kim and W. S. Kang, "Distribution Characteristics and Cost Estimation of Collection and Treatment of Deposited Marine Debris in Coastal Fisheries around the Southwestern Islands of Korea" in Journal of the Korea Society for Marine Environment & Energy, pp. 330-336, Nov. 2012. (DOI : https://doi.org/10.7846/JKOSMEE.2012.15.4.330)
  7. Y. J. Park, Y. R. Yeom and E. S. Choi, "The untold story surroundubg the sea pollution source, 'Waste phrase." Internet : http://www.ohmynews.com/NWS_Web/View/at_pg.aspx?CNTN_CD=A0002794810, Dec.15,2021[Jan.6,2022]
  8. S. Ren, K. He, R. Girshick and J. Sun, "Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks", in IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 39, No. 6, pp. 1137-1149, June. 2017. (DOI : https://doi.org/10.48550/arXiv.1506.01497)
  9. J. Redmon, S. Divvala, R. Girshick and A. Farhadi, "You Only Look Once: Unified, Real-Time Object Detection", in 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 779-788, June. 2016. (DOI : https://doi.org/10.48550/arXiv.1506.02640)
  10. Y. H. Lee and Y. S. Kim, "Comparison of CNN and YOLO for Object Detection", in Journal of the Semiconductor & Display Technology, Vol. 19, No. 1, pp. 85-92, March. 2020.
  11. Sanchez Hernandez, Sergio & Romero, H & Morales, A. (2020). A review: Comparison of performance metrics of pretrained models for object detection using the TensorFlow framework. IOP Conference Series: Materials Science and Engineering. 844. 012024. (DOI : 10.1088/1757-899X/844/1/012024)
  12. M. Tan, R. Pang, and Q. V. Le. "Efficientdet: Scalable and efficient object detection." In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 10781-10790 Jul. 2020. (DOI : 10.1109/CVPR42600.2020.01079)
  13. Ultralytics, "yolov5." Internet: https://github.com/ultralytics/yolov5, Jan.26,2020 [Oct.27,2021]
  14. NaeJoung Kwak and DongJu Kim, "A study on Detecting the Safety helmet wearing using YOLOv5-S model and transfer learning", International Journal of Advanced Culture Technology(IJACT), Vol. 10, No. 1, pp. 302-309, (2022) (DOI : https://doi.org/10.17703/IJACT.2022.10.1.302)
  15. NaeJoung Kwak and DongJu Kim, "Study On Masked Face Detection And Recognition using transfer learning", International Journal of Advanced Culture Technology(IJACT), Vol. 10, No. 1, pp. 294-301, (2022) (DOI : https://doi.org/10.17703/IJACT.2022.10.1.294)
  16. Gan Liu and Sang-Hyun Lee, "Municipal waste classification system design based on Faster-RCNN and YoloV4 mixed model", International Journal of Advanced Culture Technology(IJACT), Vol. 9, No. 3, pp. 305-314, (2021) (DOI : https://doi.org/10.17703/IJACT.2021.9.3.305)
  17. A backup space for my memory, "Regarding the view," Internet: https://nangchobi.tistory.com/474, June. 21, 2021 [Dec.2,2021]
  18. T. S. Wang, S. Y. Oh, H. S. Lee, J. W. Jang and M. Y. Kim, "A Study on the A.I Detection Model of Marine Deposition Waste Using YOLOv5" in Proceeding of the Korea Artificial-Intelligence Convergence Technology Society Vol. 1, No. 1, Pusan:PS, pp. 55, Aug. 2021.
  19. Korea Policy briefing Portal, "Measures to prevent damage to abandoned fishing areas, such as the Korea Coast Guard and ghost fisheries, etc" Internet : https://www.korea.kr/news/pressReleaseView.do?newsId=156446503, [April,14,2021]
  20. AI hub, "The image of ocean sediment waste" Internet : https://aihub.or.kr/aidata/30754, Jun.18, 2020[Oct.27,2021]
  21. J. H. Lee, Y. H. Kang and D. J. Moon "AI Data Construction for Submerged Marine Debris to Advance Marine Debris Management" in Journal of the Korea Society for Marine Environment & Energy, pp. 4-5, Oct. 2021
  22. Haneda Osamu. (2007). Knot method. Korea: Jinsun Books.