Ice Thickness From Deep Learning and Conditional Random Fields: Application to Ice-Penetrating Radar Data With Radiometric Validation
Identifying the location of the ice–bedrock interface of glaciers and ice sheets is crucial for a wide range of geophysical applications, such as searching for liquid water in basal regions and computing ice thickness to quantify ice sheet and glacier mass balance. Simple, record-by-record, approa...
Main Authors: | , , , |
---|---|
Format: | Article in Journal/Newspaper |
Language: | unknown |
Published: |
IEEE
2022
|
Subjects: | |
Online Access: | https://doi.org/10.1109/tgrs.2022.3214147 |
id |
ftcaltechauth:oai:authors.library.caltech.edu:mvad0-c2z47 |
---|---|
record_format |
openpolar |
spelling |
ftcaltechauth:oai:authors.library.caltech.edu:mvad0-c2z47 2024-06-23T07:47:55+00:00 Ice Thickness From Deep Learning and Conditional Random Fields: Application to Ice-Penetrating Radar Data With Radiometric Validation Liu-Schiaffini, Miguel Ng, Gregory Grima, Cyril Young, Duncan 2022-10-12 https://doi.org/10.1109/tgrs.2022.3214147 unknown IEEE https://doi.org/10.1109/TGRS.2022.3214147 oai:authors.library.caltech.edu:mvad0-c2z47 eprintid:117888 resolverid:CaltechAUTHORS:20221116-601553800.7 info:eu-repo/semantics/closedAccess Other IEEE Transactions on Geoscience and Remote Sensing, 60, Art. No. 5119014, (2022-10-12) General Earth and Planetary Sciences Electrical and Electronic Engineering info:eu-repo/semantics/article 2022 ftcaltechauth https://doi.org/10.1109/tgrs.2022.321414710.1109/TGRS.2022.3214147 2024-06-12T05:52:28Z Identifying the location of the ice–bedrock interface of glaciers and ice sheets is crucial for a wide range of geophysical applications, such as searching for liquid water in basal regions and computing ice thickness to quantify ice sheet and glacier mass balance. Simple, record-by-record, approaches to detecting the bottom of the ice echo may be affected by spurious off-nadir noise that requires significant manual interaction to correct. In this article, we propose a deep learning model based on convolutional neural networks (CNNs) and continuous conditional random fields (CCRFs) to automate ice bed identification and better capture fine-grained basal detail. We deploy this approach on high-capability radar sounder (HiCARS) radargrams, and this is the first time deep learning methods have been applied to this dataset. Intuitively, our CNN captures the global geometry of the ice bed, while the CCRF adjusts the initial CNN outputs to better incorporate fine-scale spatial information into the final prediction. We also develop a coherent geophysical framework using three echo characters (along-track continuity, relative delay, and signal coherency) to compare our model's outputs with those of a manually targeted approach. Our analysis suggests that our CNN + CCRF model is as suitable as the manual approach for radiometric applications, and it outperforms the manual technique in identifying the first continuous return, which is most often the near-nadir reflection. Thus, our approach is more universal than the current manual labeling methodology in the range of geophysical applications where it can be used, and it provides better confidence regarding the source of the basal return detected. This work was supported by in part by the G. Unger Vetlesen Foundation and in part by NSF under Grant 1443690 and Grant 2127606 for data analysis, and in part by NSF under Grant 0230197, Grant 0733025, and Grant 1543452; in part by the Australian Antarctic Science Project 4346; in part by the National Aeronautics and Space ... Article in Journal/Newspaper Antarc* Antarctic Ice Sheet Caltech Authors (California Institute of Technology) Antarctic |
institution |
Open Polar |
collection |
Caltech Authors (California Institute of Technology) |
op_collection_id |
ftcaltechauth |
language |
unknown |
topic |
General Earth and Planetary Sciences Electrical and Electronic Engineering |
spellingShingle |
General Earth and Planetary Sciences Electrical and Electronic Engineering Liu-Schiaffini, Miguel Ng, Gregory Grima, Cyril Young, Duncan Ice Thickness From Deep Learning and Conditional Random Fields: Application to Ice-Penetrating Radar Data With Radiometric Validation |
topic_facet |
General Earth and Planetary Sciences Electrical and Electronic Engineering |
description |
Identifying the location of the ice–bedrock interface of glaciers and ice sheets is crucial for a wide range of geophysical applications, such as searching for liquid water in basal regions and computing ice thickness to quantify ice sheet and glacier mass balance. Simple, record-by-record, approaches to detecting the bottom of the ice echo may be affected by spurious off-nadir noise that requires significant manual interaction to correct. In this article, we propose a deep learning model based on convolutional neural networks (CNNs) and continuous conditional random fields (CCRFs) to automate ice bed identification and better capture fine-grained basal detail. We deploy this approach on high-capability radar sounder (HiCARS) radargrams, and this is the first time deep learning methods have been applied to this dataset. Intuitively, our CNN captures the global geometry of the ice bed, while the CCRF adjusts the initial CNN outputs to better incorporate fine-scale spatial information into the final prediction. We also develop a coherent geophysical framework using three echo characters (along-track continuity, relative delay, and signal coherency) to compare our model's outputs with those of a manually targeted approach. Our analysis suggests that our CNN + CCRF model is as suitable as the manual approach for radiometric applications, and it outperforms the manual technique in identifying the first continuous return, which is most often the near-nadir reflection. Thus, our approach is more universal than the current manual labeling methodology in the range of geophysical applications where it can be used, and it provides better confidence regarding the source of the basal return detected. This work was supported by in part by the G. Unger Vetlesen Foundation and in part by NSF under Grant 1443690 and Grant 2127606 for data analysis, and in part by NSF under Grant 0230197, Grant 0733025, and Grant 1543452; in part by the Australian Antarctic Science Project 4346; in part by the National Aeronautics and Space ... |
format |
Article in Journal/Newspaper |
author |
Liu-Schiaffini, Miguel Ng, Gregory Grima, Cyril Young, Duncan |
author_facet |
Liu-Schiaffini, Miguel Ng, Gregory Grima, Cyril Young, Duncan |
author_sort |
Liu-Schiaffini, Miguel |
title |
Ice Thickness From Deep Learning and Conditional Random Fields: Application to Ice-Penetrating Radar Data With Radiometric Validation |
title_short |
Ice Thickness From Deep Learning and Conditional Random Fields: Application to Ice-Penetrating Radar Data With Radiometric Validation |
title_full |
Ice Thickness From Deep Learning and Conditional Random Fields: Application to Ice-Penetrating Radar Data With Radiometric Validation |
title_fullStr |
Ice Thickness From Deep Learning and Conditional Random Fields: Application to Ice-Penetrating Radar Data With Radiometric Validation |
title_full_unstemmed |
Ice Thickness From Deep Learning and Conditional Random Fields: Application to Ice-Penetrating Radar Data With Radiometric Validation |
title_sort |
ice thickness from deep learning and conditional random fields: application to ice-penetrating radar data with radiometric validation |
publisher |
IEEE |
publishDate |
2022 |
url |
https://doi.org/10.1109/tgrs.2022.3214147 |
geographic |
Antarctic |
geographic_facet |
Antarctic |
genre |
Antarc* Antarctic Ice Sheet |
genre_facet |
Antarc* Antarctic Ice Sheet |
op_source |
IEEE Transactions on Geoscience and Remote Sensing, 60, Art. No. 5119014, (2022-10-12) |
op_relation |
https://doi.org/10.1109/TGRS.2022.3214147 oai:authors.library.caltech.edu:mvad0-c2z47 eprintid:117888 resolverid:CaltechAUTHORS:20221116-601553800.7 |
op_rights |
info:eu-repo/semantics/closedAccess Other |
op_doi |
https://doi.org/10.1109/tgrs.2022.321414710.1109/TGRS.2022.3214147 |
_version_ |
1802638159337488384 |