Skip to main content

ORIGINAL RESEARCH article

Front. Remote Sens.
Sec. Microwave Remote Sensing
Volume 5 - 2024 | doi: 10.3389/frsen.2024.1417417
This article is part of the Research Topic Recent Developments of Polar Sea Ice Monitoring Using Remote Sensing Data View all articles

ARISGAN: Extreme Super-Resolution of Arctic Surface Imagery using Generative Adversarial Networks

Provisionally accepted
Christian Au-Boehm Christian Au-Boehm 1Michel Tsamados Michel Tsamados 2*Petru Manescu Petru Manescu 1So Takao So Takao 1,3
  • 1 Department of Computer Science, Faculty of Engineering Sciences, University College London, London, England, United Kingdom
  • 2 Centre for Polar Observation and Modelling, Department of Earth Sciences, Faculty of Mathematical and Physical Sciences, University College London, London, United Kingdom
  • 3 Department of Computing + Mathematical Sciences, Division of Engineering and Applied Science, California Institute of Technology, Pasadena, California, United States

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

    This research explores the application of generative artificial intelligence, specifically the novel ARISGAN framework, for generating high-resolution synthetic satellite imagery in the challenging arctic environment. The study addresses the crucial need for realistic and high-resolution surface imagery in the Arctic, vital for applications ranging from satellite retrieval systems to the well-being and safety of Inuit populations relying on detailed surface observations. The ARISGAN framework combines dense block, multireceptive field, and Pix2Pix architecture, showcasing promising results that surpass existing state-of-the-art models across diverse tasks and metrics. Land-based imagery super-resolution exhibits superior metrics in comparison to sea ice-based imagery across multiple models. This research contributes to the advancement of Earth Observation in polar regions by introducing a framework that combines advanced image processing techniques and a well-designed architecture. The ARISGAN framework's effectiveness in outperforming existing models underscores its potential for generating perceptually valid high-resolution arctic surface imagery. The study concludes with a discussion on identified limitations and proposes avenues for future research, emphasizing the importance of addressing challenges in temporal synchronicity, multi-spectral image analysis, pre-processing, and quality metrics. The findings encourage further refinement of the ARISGAN framework, ultimately advancing the quality and availability of high-resolution satellite imagery in the Arctic.

    Keywords: super-resolution, remote sensing, Computer Vision, Synthetic satellite imagery, Arctic environment, sea ice, Generative Adversarial Networks. (Min

    Received: 14 Apr 2024; Accepted: 27 Jun 2024.

    Copyright: © 2024 Au-Boehm, Tsamados, Manescu and Takao. 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: Michel Tsamados, Centre for Polar Observation and Modelling, Department of Earth Sciences, Faculty of Mathematical and Physical Sciences, University College London, London, United Kingdom

    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.