Bathymetry-guided multi-source remote sensing image domain adaptive coral reef benthic habitat classification
Coral reefs are among the most complex and biodiverse ecosystems on Earth, providing immense ecological, economic, and social value. They offer critical habitats, support fisheries and tourism, and play a key role in coastal protection. However, global climate change, overfishing, and pollution have...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Taylor & Francis Group
2025-12-01
|
| Series: | GIScience & Remote Sensing |
| Subjects: | |
| Online Access: | https://www.tandfonline.com/doi/10.1080/15481603.2025.2471193 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849763336241020928 |
|---|---|
| author | Hui Chen Liang Cheng Ka Zhang |
| author_facet | Hui Chen Liang Cheng Ka Zhang |
| author_sort | Hui Chen |
| collection | DOAJ |
| description | Coral reefs are among the most complex and biodiverse ecosystems on Earth, providing immense ecological, economic, and social value. They offer critical habitats, support fisheries and tourism, and play a key role in coastal protection. However, global climate change, overfishing, and pollution have severely threatened their survival. Therefore, accurate classification and monitoring of coral reef benthic habitats (CRBH) are essential for understanding ecosystem health and developing effective conservation strategies. Due to the complexity of marine environments, multi-source remote sensing images often exhibit significant domain shifts caused by differences in imaging conditions and platform orientations, posing challenges to the generalization of substrate classification models. To address this issue, we propose a computationally efficient Bathymetry-Guided Domain Adaptation (BGDA) method. By leveraging the domain-invariant properties of bathymetric data, the BGDA method learns the styles and cross-domain spatial relationships of CRBH under bathymetric guidance, thereby achieving cross-domain consistency and significantly enhancing the performance of unsupervised domain-adaptive substrate classification. A domain-adaptive CRBH classification dataset was constructed using Sentinel-2, PlanetScope, and Gaofen-2 satellite images. Four domain-adaptation tasks demonstrated that the BGDA method achieved a mean Intersection over Union (mIoU) of over 80%, surpassing 90% of single-domain classification performance. Additionally, the BGDA method successfully classified CRBH across 50 coral reefs in the Nansha Islands. Model transfer experiments further validated the global generalization capability of the proposed method. BGDA offers a scalable, cost-effective solution for large-scale, high-frequency CRBH classification, supporting international efforts in coral reef monitoring and conservation. |
| format | Article |
| id | doaj-art-64353e40d75d4247ba9730e543784fd5 |
| institution | DOAJ |
| issn | 1548-1603 1943-7226 |
| language | English |
| publishDate | 2025-12-01 |
| publisher | Taylor & Francis Group |
| record_format | Article |
| series | GIScience & Remote Sensing |
| spelling | doaj-art-64353e40d75d4247ba9730e543784fd52025-08-20T03:05:26ZengTaylor & Francis GroupGIScience & Remote Sensing1548-16031943-72262025-12-0162110.1080/15481603.2025.2471193Bathymetry-guided multi-source remote sensing image domain adaptive coral reef benthic habitat classificationHui Chen0Liang Cheng1Ka Zhang2School of Marine Science and Engineering, Nanjing Normal University, Nanjing, ChinaJiangsu Provincial Key Laboratory of Geographic Information Science and Technology, Nanjing University, Nanjing, ChinaKey Laboratory of Virtual Geographic Environment, Ministry of Education, Nanjing Normal University, Nanjing, ChinaCoral reefs are among the most complex and biodiverse ecosystems on Earth, providing immense ecological, economic, and social value. They offer critical habitats, support fisheries and tourism, and play a key role in coastal protection. However, global climate change, overfishing, and pollution have severely threatened their survival. Therefore, accurate classification and monitoring of coral reef benthic habitats (CRBH) are essential for understanding ecosystem health and developing effective conservation strategies. Due to the complexity of marine environments, multi-source remote sensing images often exhibit significant domain shifts caused by differences in imaging conditions and platform orientations, posing challenges to the generalization of substrate classification models. To address this issue, we propose a computationally efficient Bathymetry-Guided Domain Adaptation (BGDA) method. By leveraging the domain-invariant properties of bathymetric data, the BGDA method learns the styles and cross-domain spatial relationships of CRBH under bathymetric guidance, thereby achieving cross-domain consistency and significantly enhancing the performance of unsupervised domain-adaptive substrate classification. A domain-adaptive CRBH classification dataset was constructed using Sentinel-2, PlanetScope, and Gaofen-2 satellite images. Four domain-adaptation tasks demonstrated that the BGDA method achieved a mean Intersection over Union (mIoU) of over 80%, surpassing 90% of single-domain classification performance. Additionally, the BGDA method successfully classified CRBH across 50 coral reefs in the Nansha Islands. Model transfer experiments further validated the global generalization capability of the proposed method. BGDA offers a scalable, cost-effective solution for large-scale, high-frequency CRBH classification, supporting international efforts in coral reef monitoring and conservation.https://www.tandfonline.com/doi/10.1080/15481603.2025.2471193Coral reef benthic habitat classificationdomain adaptationbathymetryremote sensingdeep learning |
| spellingShingle | Hui Chen Liang Cheng Ka Zhang Bathymetry-guided multi-source remote sensing image domain adaptive coral reef benthic habitat classification GIScience & Remote Sensing Coral reef benthic habitat classification domain adaptation bathymetry remote sensing deep learning |
| title | Bathymetry-guided multi-source remote sensing image domain adaptive coral reef benthic habitat classification |
| title_full | Bathymetry-guided multi-source remote sensing image domain adaptive coral reef benthic habitat classification |
| title_fullStr | Bathymetry-guided multi-source remote sensing image domain adaptive coral reef benthic habitat classification |
| title_full_unstemmed | Bathymetry-guided multi-source remote sensing image domain adaptive coral reef benthic habitat classification |
| title_short | Bathymetry-guided multi-source remote sensing image domain adaptive coral reef benthic habitat classification |
| title_sort | bathymetry guided multi source remote sensing image domain adaptive coral reef benthic habitat classification |
| topic | Coral reef benthic habitat classification domain adaptation bathymetry remote sensing deep learning |
| url | https://www.tandfonline.com/doi/10.1080/15481603.2025.2471193 |
| work_keys_str_mv | AT huichen bathymetryguidedmultisourceremotesensingimagedomainadaptivecoralreefbenthichabitatclassification AT liangcheng bathymetryguidedmultisourceremotesensingimagedomainadaptivecoralreefbenthichabitatclassification AT kazhang bathymetryguidedmultisourceremotesensingimagedomainadaptivecoralreefbenthichabitatclassification |