Dual-Task Network for Terrace and Ridge Extraction: Automatic Terrace Extraction via Multi-Task Learning
Terrace detection and ridge extraction from high-resolution remote sensing imagery are crucial for soil conservation and grain production on sloping land. Traditional methods use low-to-medium resolution images, missing detailed features and lacking automation. Terrace detection and ridge extraction...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2024-02-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/16/3/568 |
| _version_ | 1827354616505827328 |
|---|---|
| author | Jun Zhang Jun Zhang Xiao Huang Weixun Zhou Huyan Fu Yuyan Chen Zhenghao Zhan |
| author_facet | Jun Zhang Jun Zhang Xiao Huang Weixun Zhou Huyan Fu Yuyan Chen Zhenghao Zhan |
| author_sort | Jun Zhang |
| collection | DOAJ |
| description | Terrace detection and ridge extraction from high-resolution remote sensing imagery are crucial for soil conservation and grain production on sloping land. Traditional methods use low-to-medium resolution images, missing detailed features and lacking automation. Terrace detection and ridge extraction are closely linked, with each influencing the other’s outcomes. However, most studies address these tasks separately, overlooking their interdependence. This research introduces a cutting-edge, multi-scale, and multi-task deep learning framework, termed DTRE-Net, designed for comprehensive terrace information extraction. This framework bridges the gap between terrace detection and ridge extraction, executing them concurrently. The network incorporates residual networks, multi-scale fusion modules, and multi-scale residual correction modules to enhance the model’s robustness in feature extraction. Comprehensive evaluations against other deep learning-based semantic segmentation methods using GF-2 terraced imagery from two distinct areas were undertaken. The results revealed intersection over union (IoU) values of 85.18% and 86.09% for different terrace morphologies and 59.79% and 73.65% for ridges. Simultaneously, we have confirmed that the connectivity of results is improved when employing multi-task learning for ridge extraction compared to directly extracting ridges. These outcomes underscore DTRE-Net’s superior capability in the automation of terrace and ridge extraction relative to alternative techniques. |
| first_indexed | 2024-03-08T03:50:19Z |
| format | Article |
| id | doaj.art-f0aed38238574d0ab6a25baeed7457a7 |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-08T03:50:19Z |
| publishDate | 2024-02-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-f0aed38238574d0ab6a25baeed7457a72024-02-09T15:21:31ZengMDPI AGRemote Sensing2072-42922024-02-0116356810.3390/rs16030568Dual-Task Network for Terrace and Ridge Extraction: Automatic Terrace Extraction via Multi-Task LearningJun Zhang0Jun Zhang1Xiao Huang2Weixun Zhou3Huyan Fu4Yuyan Chen5Zhenghao Zhan6School of Earth Science, Yunnan University, Kunming 650000, ChinaSchool of Earth Science, Yunnan University, Kunming 650000, ChinaDepartment of Environmental Sciences, Emory University, Atlanta, GA 30322, USASchool of Remote Sensing & Geomatics Engineering, Nanjing University of Information Science and Technology, Nanjing 210044, ChinaSchool of Earth Science, Yunnan University, Kunming 650000, ChinaSchool of Earth Science, Yunnan University, Kunming 650000, ChinaSchool of Earth Science, Yunnan University, Kunming 650000, ChinaTerrace detection and ridge extraction from high-resolution remote sensing imagery are crucial for soil conservation and grain production on sloping land. Traditional methods use low-to-medium resolution images, missing detailed features and lacking automation. Terrace detection and ridge extraction are closely linked, with each influencing the other’s outcomes. However, most studies address these tasks separately, overlooking their interdependence. This research introduces a cutting-edge, multi-scale, and multi-task deep learning framework, termed DTRE-Net, designed for comprehensive terrace information extraction. This framework bridges the gap between terrace detection and ridge extraction, executing them concurrently. The network incorporates residual networks, multi-scale fusion modules, and multi-scale residual correction modules to enhance the model’s robustness in feature extraction. Comprehensive evaluations against other deep learning-based semantic segmentation methods using GF-2 terraced imagery from two distinct areas were undertaken. The results revealed intersection over union (IoU) values of 85.18% and 86.09% for different terrace morphologies and 59.79% and 73.65% for ridges. Simultaneously, we have confirmed that the connectivity of results is improved when employing multi-task learning for ridge extraction compared to directly extracting ridges. These outcomes underscore DTRE-Net’s superior capability in the automation of terrace and ridge extraction relative to alternative techniques.https://www.mdpi.com/2072-4292/16/3/568multi-task learningterrace information extractionneural networkshigh-resolution remote sensing images |
| spellingShingle | Jun Zhang Jun Zhang Xiao Huang Weixun Zhou Huyan Fu Yuyan Chen Zhenghao Zhan Dual-Task Network for Terrace and Ridge Extraction: Automatic Terrace Extraction via Multi-Task Learning Remote Sensing multi-task learning terrace information extraction neural networks high-resolution remote sensing images |
| title | Dual-Task Network for Terrace and Ridge Extraction: Automatic Terrace Extraction via Multi-Task Learning |
| title_full | Dual-Task Network for Terrace and Ridge Extraction: Automatic Terrace Extraction via Multi-Task Learning |
| title_fullStr | Dual-Task Network for Terrace and Ridge Extraction: Automatic Terrace Extraction via Multi-Task Learning |
| title_full_unstemmed | Dual-Task Network for Terrace and Ridge Extraction: Automatic Terrace Extraction via Multi-Task Learning |
| title_short | Dual-Task Network for Terrace and Ridge Extraction: Automatic Terrace Extraction via Multi-Task Learning |
| title_sort | dual task network for terrace and ridge extraction automatic terrace extraction via multi task learning |
| topic | multi-task learning terrace information extraction neural networks high-resolution remote sensing images |
| url | https://www.mdpi.com/2072-4292/16/3/568 |
| work_keys_str_mv | AT junzhang dualtasknetworkforterraceandridgeextractionautomaticterraceextractionviamultitasklearning AT junzhang dualtasknetworkforterraceandridgeextractionautomaticterraceextractionviamultitasklearning AT xiaohuang dualtasknetworkforterraceandridgeextractionautomaticterraceextractionviamultitasklearning AT weixunzhou dualtasknetworkforterraceandridgeextractionautomaticterraceextractionviamultitasklearning AT huyanfu dualtasknetworkforterraceandridgeextractionautomaticterraceextractionviamultitasklearning AT yuyanchen dualtasknetworkforterraceandridgeextractionautomaticterraceextractionviamultitasklearning AT zhenghaozhan dualtasknetworkforterraceandridgeextractionautomaticterraceextractionviamultitasklearning |