Physics-Constrained Three-Dimensional Swin Transformer for Gravity Data Inversion
This paper proposes a physics-constrained 3D Swin Transformer (ST) for gravity inversion. By leveraging the self-attention mechanism in 3D ST, the method effectively models global dependencies within gravity data, enabling the network to reweight features globally and focus on critical anomalous reg...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-01-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/17/1/113 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1841549027537059840 |
|---|---|
| author | Ping Yu Longran Zhou Shuai Zhou Jian Jiao Guanlin Huang Pengyu Lu |
| author_facet | Ping Yu Longran Zhou Shuai Zhou Jian Jiao Guanlin Huang Pengyu Lu |
| author_sort | Ping Yu |
| collection | DOAJ |
| description | This paper proposes a physics-constrained 3D Swin Transformer (ST) for gravity inversion. By leveraging the self-attention mechanism in 3D ST, the method effectively models global dependencies within gravity data, enabling the network to reweight features globally and focus on critical anomalous regions. Additionally, prior gradient information is integrated into the loss function, and a hierarchical weight allocation strategy is adopted to guide the model in learning boundary information of density structures and deep-seated features more effectively. Synthetic experiments demonstrate that the proposed method achieves lower model errors, better boundary alignment, and higher inversion accuracy. The approach is further validated using gravity anomaly observations from the Gonghe Basin in Qinghai, yielding reliable and precise inversion results. |
| format | Article |
| id | doaj-art-d5bc82e1639a46d9a62869cdfc2e808b |
| institution | Kabale University |
| issn | 2072-4292 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj-art-d5bc82e1639a46d9a62869cdfc2e808b2025-01-10T13:20:16ZengMDPI AGRemote Sensing2072-42922025-01-0117111310.3390/rs17010113Physics-Constrained Three-Dimensional Swin Transformer for Gravity Data InversionPing Yu0Longran Zhou1Shuai Zhou2Jian Jiao3Guanlin Huang4Pengyu Lu5College of Geo-Exploration Science and Technology, Jilin University, Changchun 130026, ChinaCollege of Geo-Exploration Science and Technology, Jilin University, Changchun 130026, ChinaCollege of Geo-Exploration Science and Technology, Jilin University, Changchun 130026, ChinaCollege of Geo-Exploration Science and Technology, Jilin University, Changchun 130026, ChinaCollege of Geo-Exploration Science and Technology, Jilin University, Changchun 130026, ChinaCollege of Geo-Exploration Science and Technology, Jilin University, Changchun 130026, ChinaThis paper proposes a physics-constrained 3D Swin Transformer (ST) for gravity inversion. By leveraging the self-attention mechanism in 3D ST, the method effectively models global dependencies within gravity data, enabling the network to reweight features globally and focus on critical anomalous regions. Additionally, prior gradient information is integrated into the loss function, and a hierarchical weight allocation strategy is adopted to guide the model in learning boundary information of density structures and deep-seated features more effectively. Synthetic experiments demonstrate that the proposed method achieves lower model errors, better boundary alignment, and higher inversion accuracy. The approach is further validated using gravity anomaly observations from the Gonghe Basin in Qinghai, yielding reliable and precise inversion results.https://www.mdpi.com/2072-4292/17/1/113gravity inversionlong-range dependenciesself-attention mechanismphysics constrained |
| spellingShingle | Ping Yu Longran Zhou Shuai Zhou Jian Jiao Guanlin Huang Pengyu Lu Physics-Constrained Three-Dimensional Swin Transformer for Gravity Data Inversion Remote Sensing gravity inversion long-range dependencies self-attention mechanism physics constrained |
| title | Physics-Constrained Three-Dimensional Swin Transformer for Gravity Data Inversion |
| title_full | Physics-Constrained Three-Dimensional Swin Transformer for Gravity Data Inversion |
| title_fullStr | Physics-Constrained Three-Dimensional Swin Transformer for Gravity Data Inversion |
| title_full_unstemmed | Physics-Constrained Three-Dimensional Swin Transformer for Gravity Data Inversion |
| title_short | Physics-Constrained Three-Dimensional Swin Transformer for Gravity Data Inversion |
| title_sort | physics constrained three dimensional swin transformer for gravity data inversion |
| topic | gravity inversion long-range dependencies self-attention mechanism physics constrained |
| url | https://www.mdpi.com/2072-4292/17/1/113 |
| work_keys_str_mv | AT pingyu physicsconstrainedthreedimensionalswintransformerforgravitydatainversion AT longranzhou physicsconstrainedthreedimensionalswintransformerforgravitydatainversion AT shuaizhou physicsconstrainedthreedimensionalswintransformerforgravitydatainversion AT jianjiao physicsconstrainedthreedimensionalswintransformerforgravitydatainversion AT guanlinhuang physicsconstrainedthreedimensionalswintransformerforgravitydatainversion AT pengyulu physicsconstrainedthreedimensionalswintransformerforgravitydatainversion |