Version 1
: Received: 10 September 2024 / Approved: 11 September 2024 / Online: 11 September 2024 (12:31:38 CEST)
How to cite:
Murashkin, D.; Spreen, G.; Huntemann, M. Pan-Arctic Lead Maps based on Sentinel-1 SAR Images with a Convolutional Neural Network. Preprints2024, 2024090905. https://doi.org/10.20944/preprints202409.0905.v1
Murashkin, D.; Spreen, G.; Huntemann, M. Pan-Arctic Lead Maps based on Sentinel-1 SAR Images with a Convolutional Neural Network. Preprints 2024, 2024090905. https://doi.org/10.20944/preprints202409.0905.v1
Murashkin, D.; Spreen, G.; Huntemann, M. Pan-Arctic Lead Maps based on Sentinel-1 SAR Images with a Convolutional Neural Network. Preprints2024, 2024090905. https://doi.org/10.20944/preprints202409.0905.v1
APA Style
Murashkin, D., Spreen, G., & Huntemann, M. (2024). Pan-Arctic Lead Maps based on Sentinel-1 SAR Images with a Convolutional Neural Network. Preprints. https://doi.org/10.20944/preprints202409.0905.v1
Chicago/Turabian Style
Murashkin, D., Gunnar Spreen and Marcus Huntemann. 2024 "Pan-Arctic Lead Maps based on Sentinel-1 SAR Images with a Convolutional Neural Network" Preprints. https://doi.org/10.20944/preprints202409.0905.v1
Abstract
The Sentinel-1 SAR Extra Wide swath mode with its 40 meters pixel size allows accurate lead mapping in Arctic sea ice areas.
We present an improved winter lead detection algorithm based on a modified U-Net convolutional neural network.
We introduce a preprocessing procedure that balances normalized radar cross section (NRCS) values between sub-swaths.
This results in more consistent Sentinel-1 SAR images as input for the classifier, which is especially important for the cross-polarization HV channel.
In turn, image consistency is essential for automatic image semantic segmentation.
We develop a new lead detection algorithm and compare it with a previously published version.
We produce pan-Arctic, 40 meters resolution lead maps and present pan-Arctic lead area fraction on a 12 km grid for January 2019.
Within different Arctic regions, the lead area fraction is stable and varies only within the expected natural variability during the one month studied here.
The average Arctic-wide lead area fraction is 2.44 % with 0.25 % standard deviation during January 2019.
The improved lead detection algorithm leads to a better lead discrimination and extends the applicability area to the entire Arctic.
Keywords
sea ice; leads; SAR; Sentinel-1; U-Net; CNN
Subject
Environmental and Earth Sciences, Remote Sensing
Copyright:
This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
