Skip to main content

ORIGINAL RESEARCH article

Front. Mar. Sci.
Sec. Ocean Observation
Volume 11 - 2024 | doi: 10.3389/fmars.2024.1480796
This article is part of the Research Topic Remote Sensing Applications in Oceanography with Deep Learning View all 5 articles

OIEDNet: Oriented Ice Eddy Detection Network Based on the Sentinel-1 Dual-polarization Data

Provisionally accepted
Jinqun Wu Jinqun Wu 1,2Yiqin Zheng Yiqin Zheng 2Tingting Wang Tingting Wang 3Chunyong Ma Chunyong Ma 1,4*Ge Chen Ge Chen 1,4*
  • 1 Frontiers Science Center for Deep Ocean Multispheres and Earth System, School of Marine Technology, Ocean University of China, Qingdao, Shandong, China
  • 2 PIESAT Information Technology Co., Ltd., Beijing, China
  • 3 Qingdao Earthquake Prevention and Disaster Reduction Center, Qingdao, China
  • 4 Laboratory for Regional Oceanography and Numerical Modeling, Laoshan Laboratory, Qingdao, Shandong, China

The final, formatted version of the article will be published soon.

    The complex convergence of cold and warm ocean currents in the Nordic Seas provides suitable conditions for the formation and development of eddies. In the Marginal Ice Zones (MIZs), ice eddies contribute to the accelerated melting of surface sea ice by facilitating vertical heat transfer, which influences the evolution of the marginal ice zone and plays an indirect role in regulating global climate. In this paper, we employed high-resolution synthetic aperture radar (SAR) satellite imagery and proposed an oriented ice eddy detection network (OIEDNet) framework to conduct automated detection and spatiotemporal analysis of ice eddies in the Nordic Seas. Firstly, a highquality RGB false-color imaging method was developed based on Sentinel-1 dual-polarization (HH+HV) Extra-Wide Swath (EW) mode products, effectively integrating denoising algorithms and image processing techniques. Secondly, an automatic ice eddy detection method based on oriented bounding boxes (OBB) was constructed to identify the ice eddy and output features such as horizontal scales, eddy centers and rotation angles. Finally, the characteristics of the detected ice eddies in the Nordic Seas during 2022-2023 were systematically analyzed. The results demonstrate that the proposed OIEDNet exhibits significant performance in ice eddy detection.

    Keywords: Synthetic Aperture Radar, dual-polarization, Ice eddy, oriented object detection, deep learning

    Received: 14 Aug 2024; Accepted: 10 Dec 2024.

    Copyright: © 2024 Wu, Zheng, Wang, Ma and Chen. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

    * Correspondence:
    Chunyong Ma, Frontiers Science Center for Deep Ocean Multispheres and Earth System, School of Marine Technology, Ocean University of China, Qingdao, Shandong, China
    Ge Chen, Frontiers Science Center for Deep Ocean Multispheres and Earth System, School of Marine Technology, Ocean University of China, Qingdao, Shandong, China

    Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.