Journal of the Korean Society of Marine Environment & Safety (해양환경안전학회지)

The Korean Society of Marine Environment and safety (해양환경안전학회)
권호 표지
DOI QR코드

DOI QR Code

A Study on the Development of YOLO-Based Maritime Object Detection System through Geometric Interpretation of Camera Images

카메라 영상의 기하학적 해석을 통한 YOLO 알고리즘 기반 해상물체탐지시스템 개발에 관한 연구

  • Received : 2022.03.24
  • Accepted : 2022.06.27
  • Published : 2022.06.30
Download PDF
⟨ Previous Next ⟩

Abstract

For autonomous ships to be commercialized and be able to navigate in coastal water, they must be able to detect maritime obstacles. One of the most common obstacles seen in coastal area are the farm buoys. In this study, a maritime object detection system was developed that detects buoys using the YOLO algorithm and visualizes the distance and bearing between buoys and the ship through geometric interpretation of camera images. After training the maritime object detection model with 1,224 pictures of buoys, the precision of the model was 89.0%, the recall was 95.0%, and the F1-score was 92.0%. Camera calibration had been conducted to calculate the distance and bearing of an object away from the camera using the obtained image coordinates and Experiment A and B were designed to verify the performance of the maritime object detection system. As a result of verifying the performance of the maritime object detection system, it can be seen that the maritime object detection system is superior to radar in its short-distance detection capability, so that it can be used as a navigational aid along with the radar.

자율운항선박이 상용화되어 연안을 항해하기 위해서는 해상의 장애물을 탐지할 수 있어야 한다. 연안에서 가장 많이 볼 수 있는 장애물 중의 하나는 양식장의 부표이다. 이에 본 연구에서는 YOLO 알고리즘을 이용하여 해상의 부표를 탐지하고, 카메라 영상의 기하학적 해석을 통해 선박으로부터 떨어진 부표의 거리와 방위를 계산하여 장애물을 시각화하는 해상물체탐지시스템을 개발하였다. 1,224장의 양식장 부표 사진으로 해양물체탐지모델을 훈련시킨 결과, 모델의 Precision은 89.0 %, Recall은 95.0 % 그리고 F1-score는 92.0 %이었다. 얻어진 영상좌표를 이용하여 카메라로부터 떨어진 물체의 거리와 방위를 계산하기 위해 카메라 캘리브레이션을 실시하고 해상물체탐지시스템의 성능을 검증하기 위해 Experiment A, B를 설계하였다. 해상물체탐지시스템의 성능을 검증한 결과 해상물체탐지시스템이 레이더보다 근거리 탐지 능력이 뛰어나서 레이더와 더불어 항행보조장비로 사용이 가능할 것으로 판단된다.

Keywords

References

  1. Bloomenthal, J. and J. Rokne(1994), Homogeneous coordinates, The Visual Computer, Vol. 11, No. 1, pp. 15-26. https://doi.org/10.1007/BF01900696
  2. Bochkovskiy, A., C. Y. Wang, and H. Y. M. Liao(2020), Yolov4: Optimal Speed and Accuracy of Object Detection, arXiv preprint arXiv:2004.10934.
  3. Hartley, R. and Z. Zisserman(2003), Multiple view Geometry in computer vision, The second Edition, Cambridge university press, pp. 151-176.
  4. Joseph, R., D. Santosh, and G. Ross(2015), You Only Look Once: Unified, Real-time Object Detection, IEEE Conference on Computer Vision and Pattern Recognition, pp. 779-788.
  5. Kim, J. H., Y. Yun, J. Kim, K. Yun, W. S. Cheong, and S. J. Kang(2019), Accurate Camera Calibration Method for Multiview Stereoscopic Image Acquisition, Journal of Broadcast Engineering, Vol. 24, No. 6, pp. 919-927. https://doi.org/10.5909/JBE.2019.24.6.919
  6. KMI(2018), Korea Maritime Institute, A Study on Environment Improvements of Aquaculture Farms, pp. 22-24.
  7. Lee, Y. H. and Y. S. Kim(2020), Comparison of CNN and YOLO for object Detection, Journal of the Semiconductor & Display Technology, Vol. 19, No. 1, pp. 85-92.
  8. Li, Q., X. Ding, X. Wang, L. Chen, J. Son, and J. Y. Song(2021), Detection and Identification of Moving Objects at Busy Traffic Road based on YOLO v4, The Journal of The Institute of Internet, Broadcasting and Communication, Vol. 21, No. 1, pp. 141-148. https://doi.org/10.7236/JIIBC.2021.21.1.141
  9. Yu, Y. K., J. M. Koo, M. S. Oh, and I. D. Yang(2016), Rail Profile Matching Method using ICP Algorithm, The Transactions of the Korean Institute of Electrical Engineers, Vol. 65, No. 5, pp. 888-894. https://doi.org/10.5370/KIEE.2016.65.5.888
  10. Zhang, Zhengyou(2000), A Flexible New Technique for Camera Calibration, IEEE Transactions on pattern analysis and machine intelligence, Vol. 22, No. 11, pp. 1330-1334. https://doi.org/10.1109/34.888718