Proof of Concept for Sea Ice Stage of Development Classification Using Deep Learning
Accurate maps of ice concentration and ice type are needed to address increased interest in commercial marine transportation through the Arctic. RADARSAT-2 SAR imagery is the primary source of data used by expert ice analysts at the Canadian Ice Service (CIS) to produce sea ice maps over the Canadia...
Published in: | Remote Sensing |
---|---|
Main Authors: | , , , , |
Format: | Text |
Language: | English |
Published: |
Multidisciplinary Digital Publishing Institute
2020
|
Subjects: | |
Online Access: | https://doi.org/10.3390/rs12152486 |
id |
ftmdpi:oai:mdpi.com:/2072-4292/12/15/2486/ |
---|---|
record_format |
openpolar |
spelling |
ftmdpi:oai:mdpi.com:/2072-4292/12/15/2486/ 2023-08-20T04:04:13+02:00 Proof of Concept for Sea Ice Stage of Development Classification Using Deep Learning Ryan Kruk M. Christopher Fuller Alexander S. Komarov Dustin Isleifson Ian Jeffrey agris 2020-08-03 application/pdf https://doi.org/10.3390/rs12152486 EN eng Multidisciplinary Digital Publishing Institute https://dx.doi.org/10.3390/rs12152486 https://creativecommons.org/licenses/by/4.0/ Remote Sensing; Volume 12; Issue 15; Pages: 2486 sea ice Arctic Canadian sea ice chart deep learning SAR RADARSAT-2 classification U-Net DenseNet Text 2020 ftmdpi https://doi.org/10.3390/rs12152486 2023-07-31T23:52:28Z Accurate maps of ice concentration and ice type are needed to address increased interest in commercial marine transportation through the Arctic. RADARSAT-2 SAR imagery is the primary source of data used by expert ice analysts at the Canadian Ice Service (CIS) to produce sea ice maps over the Canadian territory. This study serves as a proof of concept that neural networks can be used to accurately predict ice type from SAR data. Datasets of SAR images served as inputs, and CIS ice charts served as labelled outputs to train a neural network to classify sea ice type. Our results show that DenseNet achieves the highest overall classification accuracy of 94.0% including water and the highest ice classification accuracy of 91.8% on a three class dataset using a fusion of HH and HV SAR polarizations for the input samples. The 91.8% ice classification accuracy validates the premise that a neural network can be used to effectively categorize different ice types based on SAR data. Text Arctic Sea ice MDPI Open Access Publishing Arctic Remote Sensing 12 15 2486 |
institution |
Open Polar |
collection |
MDPI Open Access Publishing |
op_collection_id |
ftmdpi |
language |
English |
topic |
sea ice Arctic Canadian sea ice chart deep learning SAR RADARSAT-2 classification U-Net DenseNet |
spellingShingle |
sea ice Arctic Canadian sea ice chart deep learning SAR RADARSAT-2 classification U-Net DenseNet Ryan Kruk M. Christopher Fuller Alexander S. Komarov Dustin Isleifson Ian Jeffrey Proof of Concept for Sea Ice Stage of Development Classification Using Deep Learning |
topic_facet |
sea ice Arctic Canadian sea ice chart deep learning SAR RADARSAT-2 classification U-Net DenseNet |
description |
Accurate maps of ice concentration and ice type are needed to address increased interest in commercial marine transportation through the Arctic. RADARSAT-2 SAR imagery is the primary source of data used by expert ice analysts at the Canadian Ice Service (CIS) to produce sea ice maps over the Canadian territory. This study serves as a proof of concept that neural networks can be used to accurately predict ice type from SAR data. Datasets of SAR images served as inputs, and CIS ice charts served as labelled outputs to train a neural network to classify sea ice type. Our results show that DenseNet achieves the highest overall classification accuracy of 94.0% including water and the highest ice classification accuracy of 91.8% on a three class dataset using a fusion of HH and HV SAR polarizations for the input samples. The 91.8% ice classification accuracy validates the premise that a neural network can be used to effectively categorize different ice types based on SAR data. |
format |
Text |
author |
Ryan Kruk M. Christopher Fuller Alexander S. Komarov Dustin Isleifson Ian Jeffrey |
author_facet |
Ryan Kruk M. Christopher Fuller Alexander S. Komarov Dustin Isleifson Ian Jeffrey |
author_sort |
Ryan Kruk |
title |
Proof of Concept for Sea Ice Stage of Development Classification Using Deep Learning |
title_short |
Proof of Concept for Sea Ice Stage of Development Classification Using Deep Learning |
title_full |
Proof of Concept for Sea Ice Stage of Development Classification Using Deep Learning |
title_fullStr |
Proof of Concept for Sea Ice Stage of Development Classification Using Deep Learning |
title_full_unstemmed |
Proof of Concept for Sea Ice Stage of Development Classification Using Deep Learning |
title_sort |
proof of concept for sea ice stage of development classification using deep learning |
publisher |
Multidisciplinary Digital Publishing Institute |
publishDate |
2020 |
url |
https://doi.org/10.3390/rs12152486 |
op_coverage |
agris |
geographic |
Arctic |
geographic_facet |
Arctic |
genre |
Arctic Sea ice |
genre_facet |
Arctic Sea ice |
op_source |
Remote Sensing; Volume 12; Issue 15; Pages: 2486 |
op_relation |
https://dx.doi.org/10.3390/rs12152486 |
op_rights |
https://creativecommons.org/licenses/by/4.0/ |
op_doi |
https://doi.org/10.3390/rs12152486 |
container_title |
Remote Sensing |
container_volume |
12 |
container_issue |
15 |
container_start_page |
2486 |
_version_ |
1774714622299865088 |