Recognition of polar lows in Sentinel-1 SAR images with deep learning
In this article, we explore the possibility of detecting polar lows in C-band synthetic aperture radar (SAR) images by means of deep learning. Specifically, we introduce a novel dataset consisting of Sentinel-1 images divided into two classes, representing the presence and absence of a maritime meso...
Published in: | IEEE Transactions on Geoscience and Remote Sensing |
---|---|
Main Authors: | , |
Format: | Article in Journal/Newspaper |
Language: | English |
Published: |
IEEE
2022
|
Subjects: | |
Online Access: | https://hdl.handle.net/10037/27275 https://doi.org/10.1109/tgrs.2022.3204886 |
id |
ftunivtroemsoe:oai:munin.uit.no:10037/27275 |
---|---|
record_format |
openpolar |
spelling |
ftunivtroemsoe:oai:munin.uit.no:10037/27275 2023-05-15T18:18:36+02:00 Recognition of polar lows in Sentinel-1 SAR images with deep learning Grahn, Jakob Bianchi, Filippo Maria 2022-09-06 https://hdl.handle.net/10037/27275 https://doi.org/10.1109/tgrs.2022.3204886 eng eng IEEE IEEE Transactions on Geoscience and Remote Sensing Grahn, Bianchi. Recognition of polar lows in Sentinel-1 SAR images with deep learning. IEEE Transactions on Geoscience and Remote Sensing. 2022 FRIDAID 2069615 doi:10.1109/tgrs.2022.3204886 0196-2892 1558-0644 https://hdl.handle.net/10037/27275 openAccess Copyright 2022 The Author(s) Journal article Tidsskriftartikkel submittedVersion 2022 ftunivtroemsoe https://doi.org/10.1109/tgrs.2022.3204886 2022-11-10T00:01:31Z In this article, we explore the possibility of detecting polar lows in C-band synthetic aperture radar (SAR) images by means of deep learning. Specifically, we introduce a novel dataset consisting of Sentinel-1 images divided into two classes, representing the presence and absence of a maritime mesocyclone, respectively. The dataset is constructed using the ECMWF reanalysis version 5 (ERA5) dataset as baseline and it consists of 2004 annotated images. To our knowledge, this is the first dataset of its kind to be publicly released. The dataset is used to train a deep learning model to classify the labeled images. Evaluated on an independent test set, the model yields an F1 score of 0.95, indicating that polar lows can be consistently detected from SAR images. Interpretability techniques applied to the deep learning model reveal that atmospheric fronts and cyclonic eyes are key features in the classification. Moreover, experimental results show that the model is accurate even if: 1) such features are significantly cropped due to the limited swath width of the SAR; 2) the features are partly covered by sea ice; and 3) land is covering significant parts of the images. By evaluating the model performance on multiple input image resolutions (pixel sizes of 500 m, 1 km, and 2 km), it is found that higher resolution yield the best performance. This emphasizes the potential of using high-resolution sensors like SAR for detecting polar lows, as compared to conventionally used sensors such as scatterometers. Article in Journal/Newspaper Sea ice University of Tromsø: Munin Open Research Archive IEEE Transactions on Geoscience and Remote Sensing 60 1 12 |
institution |
Open Polar |
collection |
University of Tromsø: Munin Open Research Archive |
op_collection_id |
ftunivtroemsoe |
language |
English |
description |
In this article, we explore the possibility of detecting polar lows in C-band synthetic aperture radar (SAR) images by means of deep learning. Specifically, we introduce a novel dataset consisting of Sentinel-1 images divided into two classes, representing the presence and absence of a maritime mesocyclone, respectively. The dataset is constructed using the ECMWF reanalysis version 5 (ERA5) dataset as baseline and it consists of 2004 annotated images. To our knowledge, this is the first dataset of its kind to be publicly released. The dataset is used to train a deep learning model to classify the labeled images. Evaluated on an independent test set, the model yields an F1 score of 0.95, indicating that polar lows can be consistently detected from SAR images. Interpretability techniques applied to the deep learning model reveal that atmospheric fronts and cyclonic eyes are key features in the classification. Moreover, experimental results show that the model is accurate even if: 1) such features are significantly cropped due to the limited swath width of the SAR; 2) the features are partly covered by sea ice; and 3) land is covering significant parts of the images. By evaluating the model performance on multiple input image resolutions (pixel sizes of 500 m, 1 km, and 2 km), it is found that higher resolution yield the best performance. This emphasizes the potential of using high-resolution sensors like SAR for detecting polar lows, as compared to conventionally used sensors such as scatterometers. |
format |
Article in Journal/Newspaper |
author |
Grahn, Jakob Bianchi, Filippo Maria |
spellingShingle |
Grahn, Jakob Bianchi, Filippo Maria Recognition of polar lows in Sentinel-1 SAR images with deep learning |
author_facet |
Grahn, Jakob Bianchi, Filippo Maria |
author_sort |
Grahn, Jakob |
title |
Recognition of polar lows in Sentinel-1 SAR images with deep learning |
title_short |
Recognition of polar lows in Sentinel-1 SAR images with deep learning |
title_full |
Recognition of polar lows in Sentinel-1 SAR images with deep learning |
title_fullStr |
Recognition of polar lows in Sentinel-1 SAR images with deep learning |
title_full_unstemmed |
Recognition of polar lows in Sentinel-1 SAR images with deep learning |
title_sort |
recognition of polar lows in sentinel-1 sar images with deep learning |
publisher |
IEEE |
publishDate |
2022 |
url |
https://hdl.handle.net/10037/27275 https://doi.org/10.1109/tgrs.2022.3204886 |
genre |
Sea ice |
genre_facet |
Sea ice |
op_relation |
IEEE Transactions on Geoscience and Remote Sensing Grahn, Bianchi. Recognition of polar lows in Sentinel-1 SAR images with deep learning. IEEE Transactions on Geoscience and Remote Sensing. 2022 FRIDAID 2069615 doi:10.1109/tgrs.2022.3204886 0196-2892 1558-0644 https://hdl.handle.net/10037/27275 |
op_rights |
openAccess Copyright 2022 The Author(s) |
op_doi |
https://doi.org/10.1109/tgrs.2022.3204886 |
container_title |
IEEE Transactions on Geoscience and Remote Sensing |
container_volume |
60 |
container_start_page |
1 |
op_container_end_page |
12 |
_version_ |
1766195227016560640 |