Deep learning based automatic grounding line delineation in DInSAR interferograms
The regular and robust mapping of grounding lines is essential for various applications related to the mass balance of marine ice sheets and glaciers, especially in Antarctica and Greenland. With Differential Interferometric Synthetic Aperture Radar (DInSAR) interferograms, it is possible to accurat...
Main Authors: | , , , |
---|---|
Format: | Text |
Language: | English |
Published: |
2024
|
Subjects: | |
Online Access: | https://doi.org/10.5194/egusphere-2024-223 https://egusphere.copernicus.org/preprints/2024/egusphere-2024-223/ |
id |
ftcopernicus:oai:publications.copernicus.org:egusphere117735 |
---|---|
record_format |
openpolar |
spelling |
ftcopernicus:oai:publications.copernicus.org:egusphere117735 2024-09-15T17:46:56+00:00 Deep learning based automatic grounding line delineation in DInSAR interferograms Ramanath Tarekere, Sindhu Krieger, Lukas Floricioiu, Dana Heidler, Konrad 2024-03-11 application/pdf https://doi.org/10.5194/egusphere-2024-223 https://egusphere.copernicus.org/preprints/2024/egusphere-2024-223/ eng eng doi:10.5194/egusphere-2024-223 https://egusphere.copernicus.org/preprints/2024/egusphere-2024-223/ eISSN: Text 2024 ftcopernicus https://doi.org/10.5194/egusphere-2024-223 2024-08-28T05:24:15Z The regular and robust mapping of grounding lines is essential for various applications related to the mass balance of marine ice sheets and glaciers, especially in Antarctica and Greenland. With Differential Interferometric Synthetic Aperture Radar (DInSAR) interferograms, it is possible to accurately capture the tide-induced bending of the ice shelf at a continent-wide scale and a temporal resolution of a few days. While current processing chains typically automatically generate differential interferograms, grounding lines are still primarily identified and delineated on the interferograms by a human operator. This method is time-consuming and inefficient, considering the volume of data from current and future SAR missions. We developed a pipeline that utilizes the Holistically-Nested Edge Detection (HED) neural network to delineate DInSAR interferograms automatically. We trained HED in a supervised manner using 421 manually annotated grounding lines for outlet glaciers and ice shelves on the Antarctic Ice Sheet. We also assessed the contribution of non-interferometric features like elevation, ice velocity and differential tide levels towards the delineation task. Our best-performing network generated grounding lines with a median distance of 186 m from the manual delineations. Additionally, we applied the network to generate grounding lines for undelineated interferograms, demonstrating the network's generalization capabilities and potential to generate high-resolution temporal and spatial mappings. Text Antarc* Antarctic Antarctica Greenland Ice Sheet Ice Shelf Ice Shelves Copernicus Publications: E-Journals |
institution |
Open Polar |
collection |
Copernicus Publications: E-Journals |
op_collection_id |
ftcopernicus |
language |
English |
description |
The regular and robust mapping of grounding lines is essential for various applications related to the mass balance of marine ice sheets and glaciers, especially in Antarctica and Greenland. With Differential Interferometric Synthetic Aperture Radar (DInSAR) interferograms, it is possible to accurately capture the tide-induced bending of the ice shelf at a continent-wide scale and a temporal resolution of a few days. While current processing chains typically automatically generate differential interferograms, grounding lines are still primarily identified and delineated on the interferograms by a human operator. This method is time-consuming and inefficient, considering the volume of data from current and future SAR missions. We developed a pipeline that utilizes the Holistically-Nested Edge Detection (HED) neural network to delineate DInSAR interferograms automatically. We trained HED in a supervised manner using 421 manually annotated grounding lines for outlet glaciers and ice shelves on the Antarctic Ice Sheet. We also assessed the contribution of non-interferometric features like elevation, ice velocity and differential tide levels towards the delineation task. Our best-performing network generated grounding lines with a median distance of 186 m from the manual delineations. Additionally, we applied the network to generate grounding lines for undelineated interferograms, demonstrating the network's generalization capabilities and potential to generate high-resolution temporal and spatial mappings. |
format |
Text |
author |
Ramanath Tarekere, Sindhu Krieger, Lukas Floricioiu, Dana Heidler, Konrad |
spellingShingle |
Ramanath Tarekere, Sindhu Krieger, Lukas Floricioiu, Dana Heidler, Konrad Deep learning based automatic grounding line delineation in DInSAR interferograms |
author_facet |
Ramanath Tarekere, Sindhu Krieger, Lukas Floricioiu, Dana Heidler, Konrad |
author_sort |
Ramanath Tarekere, Sindhu |
title |
Deep learning based automatic grounding line delineation in DInSAR interferograms |
title_short |
Deep learning based automatic grounding line delineation in DInSAR interferograms |
title_full |
Deep learning based automatic grounding line delineation in DInSAR interferograms |
title_fullStr |
Deep learning based automatic grounding line delineation in DInSAR interferograms |
title_full_unstemmed |
Deep learning based automatic grounding line delineation in DInSAR interferograms |
title_sort |
deep learning based automatic grounding line delineation in dinsar interferograms |
publishDate |
2024 |
url |
https://doi.org/10.5194/egusphere-2024-223 https://egusphere.copernicus.org/preprints/2024/egusphere-2024-223/ |
genre |
Antarc* Antarctic Antarctica Greenland Ice Sheet Ice Shelf Ice Shelves |
genre_facet |
Antarc* Antarctic Antarctica Greenland Ice Sheet Ice Shelf Ice Shelves |
op_source |
eISSN: |
op_relation |
doi:10.5194/egusphere-2024-223 https://egusphere.copernicus.org/preprints/2024/egusphere-2024-223/ |
op_doi |
https://doi.org/10.5194/egusphere-2024-223 |
_version_ |
1810495375035858944 |