Ice floe segmentation and floe size distribution in airborne and high-resolution optical satellite images: towards an automated labelling deep learning approach
Floe size distribution (FSD) has become a parameter of great interest in observations of sea ice because of its importance in affecting climate change, marine ecosystems, and human activities in the polar ocean. A most effective way to monitor FSD in the ice-covered regions is to apply image process...
Published in: | The Cryosphere |
---|---|
Main Authors: | , |
Format: | Text |
Language: | English |
Published: |
2023
|
Subjects: | |
Online Access: | https://doi.org/10.5194/tc-17-5519-2023 https://tc.copernicus.org/articles/17/5519/2023/ |
id |
ftcopernicus:oai:publications.copernicus.org:tc109756 |
---|---|
record_format |
openpolar |
spelling |
ftcopernicus:oai:publications.copernicus.org:tc109756 2024-09-15T18:35:31+00:00 Ice floe segmentation and floe size distribution in airborne and high-resolution optical satellite images: towards an automated labelling deep learning approach Zhang, Qin Hughes, Nick 2023-12-22 application/pdf https://doi.org/10.5194/tc-17-5519-2023 https://tc.copernicus.org/articles/17/5519/2023/ eng eng doi:10.5194/tc-17-5519-2023 https://tc.copernicus.org/articles/17/5519/2023/ eISSN: 1994-0424 Text 2023 ftcopernicus https://doi.org/10.5194/tc-17-5519-2023 2024-08-28T05:24:15Z Floe size distribution (FSD) has become a parameter of great interest in observations of sea ice because of its importance in affecting climate change, marine ecosystems, and human activities in the polar ocean. A most effective way to monitor FSD in the ice-covered regions is to apply image processing techniques to airborne and satellite remote sensing data, where the segmentation of individual ice floes is a challenge in obtaining FSD from remotely sensed images. In this study, we adopt a deep learning (DL) semantic segmentation network to fast and adaptive implement the task of ice floe instance segmentation. In order to alleviate the costly and time-consuming data annotation problem of model training, classical image processing technique is applied to automatically label ice floes in local-scale marginal ice zone (MIZ). Several state-of-the-art (SoA) semantic segmentation models are then trained on the labelled MIZ dataset and further applied to additional large-scale optical Sentinel-2 images to evaluate their performance in floe instance segmentation and to determine the best model. A post-processing algorithm is also proposed in our work to refine the segmentation. Our approach has been applied to both airborne and high-resolution optical (HRO) satellite images to derive acceptable FSDs at local and global scales. Text Sea ice Copernicus Publications: E-Journals The Cryosphere 17 12 5519 5537 |
institution |
Open Polar |
collection |
Copernicus Publications: E-Journals |
op_collection_id |
ftcopernicus |
language |
English |
description |
Floe size distribution (FSD) has become a parameter of great interest in observations of sea ice because of its importance in affecting climate change, marine ecosystems, and human activities in the polar ocean. A most effective way to monitor FSD in the ice-covered regions is to apply image processing techniques to airborne and satellite remote sensing data, where the segmentation of individual ice floes is a challenge in obtaining FSD from remotely sensed images. In this study, we adopt a deep learning (DL) semantic segmentation network to fast and adaptive implement the task of ice floe instance segmentation. In order to alleviate the costly and time-consuming data annotation problem of model training, classical image processing technique is applied to automatically label ice floes in local-scale marginal ice zone (MIZ). Several state-of-the-art (SoA) semantic segmentation models are then trained on the labelled MIZ dataset and further applied to additional large-scale optical Sentinel-2 images to evaluate their performance in floe instance segmentation and to determine the best model. A post-processing algorithm is also proposed in our work to refine the segmentation. Our approach has been applied to both airborne and high-resolution optical (HRO) satellite images to derive acceptable FSDs at local and global scales. |
format |
Text |
author |
Zhang, Qin Hughes, Nick |
spellingShingle |
Zhang, Qin Hughes, Nick Ice floe segmentation and floe size distribution in airborne and high-resolution optical satellite images: towards an automated labelling deep learning approach |
author_facet |
Zhang, Qin Hughes, Nick |
author_sort |
Zhang, Qin |
title |
Ice floe segmentation and floe size distribution in airborne and high-resolution optical satellite images: towards an automated labelling deep learning approach |
title_short |
Ice floe segmentation and floe size distribution in airborne and high-resolution optical satellite images: towards an automated labelling deep learning approach |
title_full |
Ice floe segmentation and floe size distribution in airborne and high-resolution optical satellite images: towards an automated labelling deep learning approach |
title_fullStr |
Ice floe segmentation and floe size distribution in airborne and high-resolution optical satellite images: towards an automated labelling deep learning approach |
title_full_unstemmed |
Ice floe segmentation and floe size distribution in airborne and high-resolution optical satellite images: towards an automated labelling deep learning approach |
title_sort |
ice floe segmentation and floe size distribution in airborne and high-resolution optical satellite images: towards an automated labelling deep learning approach |
publishDate |
2023 |
url |
https://doi.org/10.5194/tc-17-5519-2023 https://tc.copernicus.org/articles/17/5519/2023/ |
genre |
Sea ice |
genre_facet |
Sea ice |
op_source |
eISSN: 1994-0424 |
op_relation |
doi:10.5194/tc-17-5519-2023 https://tc.copernicus.org/articles/17/5519/2023/ |
op_doi |
https://doi.org/10.5194/tc-17-5519-2023 |
container_title |
The Cryosphere |
container_volume |
17 |
container_issue |
12 |
container_start_page |
5519 |
op_container_end_page |
5537 |
_version_ |
1810478701205258240 |