Improved Gravity Inversion Method Based on Deep Learning with Physical Constraint and Its Application to the Airborne Gravity Data in East Antarctica
This paper aims to solve the limitations of traditional gravity physical property inversion methods such as insufficient depth resolution and difficulties in parameter selection, by proposing an improved 3D gravity inversion method based on deep learning. The deep learning network model is establish...
Published in: | Remote Sensing |
---|---|
Main Authors: | , , , , , , |
Format: | Article in Journal/Newspaper |
Language: | English |
Published: |
MDPI AG
2023
|
Subjects: | |
Online Access: | https://doi.org/10.3390/rs15204933 https://doaj.org/article/df5019052f5a4c26883a1472074228f9 |
id |
ftdoajarticles:oai:doaj.org/article:df5019052f5a4c26883a1472074228f9 |
---|---|
record_format |
openpolar |
spelling |
ftdoajarticles:oai:doaj.org/article:df5019052f5a4c26883a1472074228f9 2023-11-12T04:06:09+01:00 Improved Gravity Inversion Method Based on Deep Learning with Physical Constraint and Its Application to the Airborne Gravity Data in East Antarctica Guochao Wu Yue Wei Siyuan Dong Tao Zhang Chunguo Yang Linjiang Qin Qingsheng Guan 2023-10-01T00:00:00Z https://doi.org/10.3390/rs15204933 https://doaj.org/article/df5019052f5a4c26883a1472074228f9 EN eng MDPI AG https://www.mdpi.com/2072-4292/15/20/4933 https://doaj.org/toc/2072-4292 doi:10.3390/rs15204933 2072-4292 https://doaj.org/article/df5019052f5a4c26883a1472074228f9 Remote Sensing, Vol 15, Iss 4933, p 4933 (2023) gravity inversion deep learning U-net network physical constraint East Antarctica Science Q article 2023 ftdoajarticles https://doi.org/10.3390/rs15204933 2023-10-29T00:35:47Z This paper aims to solve the limitations of traditional gravity physical property inversion methods such as insufficient depth resolution and difficulties in parameter selection, by proposing an improved 3D gravity inversion method based on deep learning. The deep learning network model is established using the fully convolutional U-net network. To enhance the generalization ability of the sample set, the large-scale training set and test set are generated by the random walk, based on the forward theory. Founded on the traditional loss function’s definition, this paper introduces an improvement incorporating a physical constraint to measure the degree of data fitting between the predicted and the real gravity data. This improvement significantly boosted the accuracy of the deep learning inversion method, as verified through both a single model and an intricate combination model. Finally, we applied this improved inversion method to the gravity data from the Gamburtsev Subglacial Mountains in the interior of East Antarctica, obtaining a comprehensive 3D crustal density structure. The results provide new evidence for the presence of a dense crustal root situated beneath the central Gamburtsev Province near the Gamburtsev Suture. Article in Journal/Newspaper Antarc* Antarctica East Antarctica Directory of Open Access Journals: DOAJ Articles Remote Sensing 15 20 4933 |
institution |
Open Polar |
collection |
Directory of Open Access Journals: DOAJ Articles |
op_collection_id |
ftdoajarticles |
language |
English |
topic |
gravity inversion deep learning U-net network physical constraint East Antarctica Science Q |
spellingShingle |
gravity inversion deep learning U-net network physical constraint East Antarctica Science Q Guochao Wu Yue Wei Siyuan Dong Tao Zhang Chunguo Yang Linjiang Qin Qingsheng Guan Improved Gravity Inversion Method Based on Deep Learning with Physical Constraint and Its Application to the Airborne Gravity Data in East Antarctica |
topic_facet |
gravity inversion deep learning U-net network physical constraint East Antarctica Science Q |
description |
This paper aims to solve the limitations of traditional gravity physical property inversion methods such as insufficient depth resolution and difficulties in parameter selection, by proposing an improved 3D gravity inversion method based on deep learning. The deep learning network model is established using the fully convolutional U-net network. To enhance the generalization ability of the sample set, the large-scale training set and test set are generated by the random walk, based on the forward theory. Founded on the traditional loss function’s definition, this paper introduces an improvement incorporating a physical constraint to measure the degree of data fitting between the predicted and the real gravity data. This improvement significantly boosted the accuracy of the deep learning inversion method, as verified through both a single model and an intricate combination model. Finally, we applied this improved inversion method to the gravity data from the Gamburtsev Subglacial Mountains in the interior of East Antarctica, obtaining a comprehensive 3D crustal density structure. The results provide new evidence for the presence of a dense crustal root situated beneath the central Gamburtsev Province near the Gamburtsev Suture. |
format |
Article in Journal/Newspaper |
author |
Guochao Wu Yue Wei Siyuan Dong Tao Zhang Chunguo Yang Linjiang Qin Qingsheng Guan |
author_facet |
Guochao Wu Yue Wei Siyuan Dong Tao Zhang Chunguo Yang Linjiang Qin Qingsheng Guan |
author_sort |
Guochao Wu |
title |
Improved Gravity Inversion Method Based on Deep Learning with Physical Constraint and Its Application to the Airborne Gravity Data in East Antarctica |
title_short |
Improved Gravity Inversion Method Based on Deep Learning with Physical Constraint and Its Application to the Airborne Gravity Data in East Antarctica |
title_full |
Improved Gravity Inversion Method Based on Deep Learning with Physical Constraint and Its Application to the Airborne Gravity Data in East Antarctica |
title_fullStr |
Improved Gravity Inversion Method Based on Deep Learning with Physical Constraint and Its Application to the Airborne Gravity Data in East Antarctica |
title_full_unstemmed |
Improved Gravity Inversion Method Based on Deep Learning with Physical Constraint and Its Application to the Airborne Gravity Data in East Antarctica |
title_sort |
improved gravity inversion method based on deep learning with physical constraint and its application to the airborne gravity data in east antarctica |
publisher |
MDPI AG |
publishDate |
2023 |
url |
https://doi.org/10.3390/rs15204933 https://doaj.org/article/df5019052f5a4c26883a1472074228f9 |
genre |
Antarc* Antarctica East Antarctica |
genre_facet |
Antarc* Antarctica East Antarctica |
op_source |
Remote Sensing, Vol 15, Iss 4933, p 4933 (2023) |
op_relation |
https://www.mdpi.com/2072-4292/15/20/4933 https://doaj.org/toc/2072-4292 doi:10.3390/rs15204933 2072-4292 https://doaj.org/article/df5019052f5a4c26883a1472074228f9 |
op_doi |
https://doi.org/10.3390/rs15204933 |
container_title |
Remote Sensing |
container_volume |
15 |
container_issue |
20 |
container_start_page |
4933 |
_version_ |
1782327267607707648 |