Recognizing the Shape and Size of Tundra Lakes in Synthetic Aperture Radar (SAR) Images Using Deep Learning Segmentation
Permafrost tundra contains more than twice as much carbon as is currently in the atmosphere, and it is warming six times as fast as the global mean. Tundra lakes dynamics is a robust indicator of global climate processes, and is still not well understood. Satellite data, particularly, from synthetic...
Published in: | Remote Sensing |
---|---|
Main Authors: | , , , , , |
Format: | Text |
Language: | English |
Published: |
Multidisciplinary Digital Publishing Institute
2023
|
Subjects: | |
Online Access: | https://doi.org/10.3390/rs15051298 |
id |
ftmdpi:oai:mdpi.com:/2072-4292/15/5/1298/ |
---|---|
record_format |
openpolar |
spelling |
ftmdpi:oai:mdpi.com:/2072-4292/15/5/1298/ 2023-08-20T04:04:50+02:00 Recognizing the Shape and Size of Tundra Lakes in Synthetic Aperture Radar (SAR) Images Using Deep Learning Segmentation Denis Demchev Ivan Sudakow Alexander Khodos Irina Abramova Dmitry Lyakhov Dominik Michels agris 2023-02-26 application/pdf https://doi.org/10.3390/rs15051298 EN eng Multidisciplinary Digital Publishing Institute Environmental Remote Sensing https://dx.doi.org/10.3390/rs15051298 https://creativecommons.org/licenses/by/4.0/ Remote Sensing; Volume 15; Issue 5; Pages: 1298 tundra lakes synthetic aperture radar Sentinel-1 U-Net size distribution Arctic climate Text 2023 ftmdpi https://doi.org/10.3390/rs15051298 2023-08-01T09:00:21Z Permafrost tundra contains more than twice as much carbon as is currently in the atmosphere, and it is warming six times as fast as the global mean. Tundra lakes dynamics is a robust indicator of global climate processes, and is still not well understood. Satellite data, particularly, from synthetic aperture radar (SAR) is a suitable tool for tundra lakes recognition and monitoring of their changes. However, manual analysis of lake boundaries can be slow and inefficient; therefore, reliable automated algorithms are required. To address this issue, we propose a two-stage approach, comprising instance deep-learning-based segmentation by U-Net, followed by semantic segmentation based on a watershed algorithm for separating touching and overlapping lakes. Implementation of this concept is essential for accurate sizes and shapes estimation of an individual lake. Here, we evaluated the performance of the proposed approach on lakes, manually extracted from tens of C-band SAR images from Sentinel-1, which were collected in the Yamal Peninsula and Alaska areas in the summer months of 2015–2022. An accuracy of 0.73, in terms of the Jaccard similarity index, was achieved. The lake’s perimeter, area and fractal sizes were estimated, based on the algorithm framework output from hundreds of SAR images. It was recognized as lognormal distributed. The evaluation of the results indicated the efficiency of the proposed approach for accurate automatic estimation of tundra lake shapes and sizes, and its potential to be used for further studies on tundra lake dynamics, in the context of global climate change, aimed at revealing new factors that could cause the planet to warm or cool. Text Arctic Climate change permafrost Tundra Yamal Peninsula Alaska MDPI Open Access Publishing Arctic Yamal Peninsula ENVELOPE(69.873,69.873,70.816,70.816) Remote Sensing 15 5 1298 |
institution |
Open Polar |
collection |
MDPI Open Access Publishing |
op_collection_id |
ftmdpi |
language |
English |
topic |
tundra lakes synthetic aperture radar Sentinel-1 U-Net size distribution Arctic climate |
spellingShingle |
tundra lakes synthetic aperture radar Sentinel-1 U-Net size distribution Arctic climate Denis Demchev Ivan Sudakow Alexander Khodos Irina Abramova Dmitry Lyakhov Dominik Michels Recognizing the Shape and Size of Tundra Lakes in Synthetic Aperture Radar (SAR) Images Using Deep Learning Segmentation |
topic_facet |
tundra lakes synthetic aperture radar Sentinel-1 U-Net size distribution Arctic climate |
description |
Permafrost tundra contains more than twice as much carbon as is currently in the atmosphere, and it is warming six times as fast as the global mean. Tundra lakes dynamics is a robust indicator of global climate processes, and is still not well understood. Satellite data, particularly, from synthetic aperture radar (SAR) is a suitable tool for tundra lakes recognition and monitoring of their changes. However, manual analysis of lake boundaries can be slow and inefficient; therefore, reliable automated algorithms are required. To address this issue, we propose a two-stage approach, comprising instance deep-learning-based segmentation by U-Net, followed by semantic segmentation based on a watershed algorithm for separating touching and overlapping lakes. Implementation of this concept is essential for accurate sizes and shapes estimation of an individual lake. Here, we evaluated the performance of the proposed approach on lakes, manually extracted from tens of C-band SAR images from Sentinel-1, which were collected in the Yamal Peninsula and Alaska areas in the summer months of 2015–2022. An accuracy of 0.73, in terms of the Jaccard similarity index, was achieved. The lake’s perimeter, area and fractal sizes were estimated, based on the algorithm framework output from hundreds of SAR images. It was recognized as lognormal distributed. The evaluation of the results indicated the efficiency of the proposed approach for accurate automatic estimation of tundra lake shapes and sizes, and its potential to be used for further studies on tundra lake dynamics, in the context of global climate change, aimed at revealing new factors that could cause the planet to warm or cool. |
format |
Text |
author |
Denis Demchev Ivan Sudakow Alexander Khodos Irina Abramova Dmitry Lyakhov Dominik Michels |
author_facet |
Denis Demchev Ivan Sudakow Alexander Khodos Irina Abramova Dmitry Lyakhov Dominik Michels |
author_sort |
Denis Demchev |
title |
Recognizing the Shape and Size of Tundra Lakes in Synthetic Aperture Radar (SAR) Images Using Deep Learning Segmentation |
title_short |
Recognizing the Shape and Size of Tundra Lakes in Synthetic Aperture Radar (SAR) Images Using Deep Learning Segmentation |
title_full |
Recognizing the Shape and Size of Tundra Lakes in Synthetic Aperture Radar (SAR) Images Using Deep Learning Segmentation |
title_fullStr |
Recognizing the Shape and Size of Tundra Lakes in Synthetic Aperture Radar (SAR) Images Using Deep Learning Segmentation |
title_full_unstemmed |
Recognizing the Shape and Size of Tundra Lakes in Synthetic Aperture Radar (SAR) Images Using Deep Learning Segmentation |
title_sort |
recognizing the shape and size of tundra lakes in synthetic aperture radar (sar) images using deep learning segmentation |
publisher |
Multidisciplinary Digital Publishing Institute |
publishDate |
2023 |
url |
https://doi.org/10.3390/rs15051298 |
op_coverage |
agris |
long_lat |
ENVELOPE(69.873,69.873,70.816,70.816) |
geographic |
Arctic Yamal Peninsula |
geographic_facet |
Arctic Yamal Peninsula |
genre |
Arctic Climate change permafrost Tundra Yamal Peninsula Alaska |
genre_facet |
Arctic Climate change permafrost Tundra Yamal Peninsula Alaska |
op_source |
Remote Sensing; Volume 15; Issue 5; Pages: 1298 |
op_relation |
Environmental Remote Sensing https://dx.doi.org/10.3390/rs15051298 |
op_rights |
https://creativecommons.org/licenses/by/4.0/ |
op_doi |
https://doi.org/10.3390/rs15051298 |
container_title |
Remote Sensing |
container_volume |
15 |
container_issue |
5 |
container_start_page |
1298 |
_version_ |
1774715255797055488 |