Fine-grained wetland classification for national wetland reserves using multi-source remote sensing data and Pixel Information Expert Engine (PIE-Engine)
ABSTRACTTimely and accurate wetland information is necessary for wetland resource management. Recent advances in machine learning and remote sensing have facilitated cost-effective monitoring of wetlands. However, reliable methods for fine-grained and rapid wetland mapping are still lacking. To addr...
| Main Authors: | , , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Taylor & Francis Group
2023-12-01
|
| Series: | GIScience & Remote Sensing |
| Subjects: | |
| Online Access: | https://www.tandfonline.com/doi/10.1080/15481603.2023.2286746 |
| _version_ | 1827632294122225664 |
|---|---|
| author | Han Liu Tongkui Liao Yu Wang Xiaoming Qian Xiaochen Liu Chengming Li Shiwei Li Zhanlei Guan Lijue Zhu Xiaoyuan Zhou Chong Liu Tengyun Hu Ming Luo |
| author_facet | Han Liu Tongkui Liao Yu Wang Xiaoming Qian Xiaochen Liu Chengming Li Shiwei Li Zhanlei Guan Lijue Zhu Xiaoyuan Zhou Chong Liu Tengyun Hu Ming Luo |
| author_sort | Han Liu |
| collection | DOAJ |
| description | ABSTRACTTimely and accurate wetland information is necessary for wetland resource management. Recent advances in machine learning and remote sensing have facilitated cost-effective monitoring of wetlands. However, reliable methods for fine-grained and rapid wetland mapping are still lacking. To address the issue, a wetland sample set with 20 categories for China was collected based on a sampling strategy that combines automatic sample generation and visual interpretation. Simultaneously, a novel multi-stage method for fine-grained wetland classification was proposed, which integrates pixel-based and object-based strategies using ensemble learning algorithms and multi-source remote sensing data. First, a pixel-based ensemble learning algorithm was implemented to classify five rough wetland categories and six non-wetland categories. Second, an object-based ensemble learning approach was designed to separate the water cover in the pixel-based classification results into eight detailed categories. Third, the merged pixel-based and object-based classification results were refined with knowledge-based post-processing procedures to identify 14 fine-grained wetland categories. Results using the Pixel Information Expert Engine (PIE-Engine) cloud platform proved the effectiveness of the proposed wetland classification method. The overall accuracy, kappa, and weighted F1 reached 87.39%, 82.80%, and 86.02%, respectively. The adopted ensemble learning algorithm yielded better performance than classifiers such as CatBoost, random forest, and XGBoost. The incorporation of spectral, texture, shape, topographic, and geographic features from multi-source data contributed to differentiating wetland categories. According to the relative contribution, spectral indexes (NDVI and NDWI), texture features (sum average and contrast), and topographic features (slope and elevation) were identified as important leading predictors for the first-stage pixel-based classification. Shape features (shape index and compactness) and auxiliary features (geographic location) were crucial predictors for the second-stage object-based classification. Compared with other products, our 10-m wetland mapping results for national wetland reserves were rich in detail and fine in categories. Overall, the constructed sample set and developed classification method show promise in laying a foundation for large-scale wetland mapping. The derived wetland maps can provide support for wetland protection and restoration. |
| first_indexed | 2024-03-09T14:34:04Z |
| format | Article |
| id | doaj.art-c5fb333aa3714c77bb13a1ad6f67ca94 |
| institution | Directory Open Access Journal |
| issn | 1548-1603 1943-7226 |
| language | English |
| last_indexed | 2024-03-09T14:34:04Z |
| publishDate | 2023-12-01 |
| publisher | Taylor & Francis Group |
| record_format | Article |
| series | GIScience & Remote Sensing |
| spelling | doaj.art-c5fb333aa3714c77bb13a1ad6f67ca942023-11-27T16:01:39ZengTaylor & Francis GroupGIScience & Remote Sensing1548-16031943-72262023-12-0160110.1080/15481603.2023.2286746Fine-grained wetland classification for national wetland reserves using multi-source remote sensing data and Pixel Information Expert Engine (PIE-Engine)Han Liu0Tongkui Liao1Yu Wang2Xiaoming Qian3Xiaochen Liu4Chengming Li5Shiwei Li6Zhanlei Guan7Lijue Zhu8Xiaoyuan Zhou9Chong Liu10Tengyun Hu11Ming Luo12Key Laboratory of Land Consolidation and Rehabilitation, Technology Innovation Center for Land Engineering, Land Consolidation and Rehabilitation Center, Ministry of Natural Resources, Beijing, ChinaPiesat Information Technology Co, Ltd, Beijing, ChinaNational Disaster Reduction Center of China, Ministry of Emergency Management, Beijing, ChinaPiesat Information Technology Co, Ltd, Beijing, ChinaPiesat Information Technology Co, Ltd, Beijing, ChinaPiesat Information Technology Co, Ltd, Beijing, ChinaPiesat Information Technology Co, Ltd, Beijing, ChinaPiesat Information Technology Co, Ltd, Beijing, ChinaPiesat Information Technology Co, Ltd, Beijing, ChinaPiesat Information Technology Co, Ltd, Beijing, ChinaPiesat Information Technology Co, Ltd, Beijing, ChinaBeijing Municipal Institute of City Planning and Design, Beijing, ChinaKey Laboratory of Land Consolidation and Rehabilitation, Technology Innovation Center for Land Engineering, Land Consolidation and Rehabilitation Center, Ministry of Natural Resources, Beijing, ChinaABSTRACTTimely and accurate wetland information is necessary for wetland resource management. Recent advances in machine learning and remote sensing have facilitated cost-effective monitoring of wetlands. However, reliable methods for fine-grained and rapid wetland mapping are still lacking. To address the issue, a wetland sample set with 20 categories for China was collected based on a sampling strategy that combines automatic sample generation and visual interpretation. Simultaneously, a novel multi-stage method for fine-grained wetland classification was proposed, which integrates pixel-based and object-based strategies using ensemble learning algorithms and multi-source remote sensing data. First, a pixel-based ensemble learning algorithm was implemented to classify five rough wetland categories and six non-wetland categories. Second, an object-based ensemble learning approach was designed to separate the water cover in the pixel-based classification results into eight detailed categories. Third, the merged pixel-based and object-based classification results were refined with knowledge-based post-processing procedures to identify 14 fine-grained wetland categories. Results using the Pixel Information Expert Engine (PIE-Engine) cloud platform proved the effectiveness of the proposed wetland classification method. The overall accuracy, kappa, and weighted F1 reached 87.39%, 82.80%, and 86.02%, respectively. The adopted ensemble learning algorithm yielded better performance than classifiers such as CatBoost, random forest, and XGBoost. The incorporation of spectral, texture, shape, topographic, and geographic features from multi-source data contributed to differentiating wetland categories. According to the relative contribution, spectral indexes (NDVI and NDWI), texture features (sum average and contrast), and topographic features (slope and elevation) were identified as important leading predictors for the first-stage pixel-based classification. Shape features (shape index and compactness) and auxiliary features (geographic location) were crucial predictors for the second-stage object-based classification. Compared with other products, our 10-m wetland mapping results for national wetland reserves were rich in detail and fine in categories. Overall, the constructed sample set and developed classification method show promise in laying a foundation for large-scale wetland mapping. The derived wetland maps can provide support for wetland protection and restoration.https://www.tandfonline.com/doi/10.1080/15481603.2023.2286746Wetland mappingremote sensingensemble learningpixel-based classificationobject-based classificationPIE-Engine |
| spellingShingle | Han Liu Tongkui Liao Yu Wang Xiaoming Qian Xiaochen Liu Chengming Li Shiwei Li Zhanlei Guan Lijue Zhu Xiaoyuan Zhou Chong Liu Tengyun Hu Ming Luo Fine-grained wetland classification for national wetland reserves using multi-source remote sensing data and Pixel Information Expert Engine (PIE-Engine) GIScience & Remote Sensing Wetland mapping remote sensing ensemble learning pixel-based classification object-based classification PIE-Engine |
| title | Fine-grained wetland classification for national wetland reserves using multi-source remote sensing data and Pixel Information Expert Engine (PIE-Engine) |
| title_full | Fine-grained wetland classification for national wetland reserves using multi-source remote sensing data and Pixel Information Expert Engine (PIE-Engine) |
| title_fullStr | Fine-grained wetland classification for national wetland reserves using multi-source remote sensing data and Pixel Information Expert Engine (PIE-Engine) |
| title_full_unstemmed | Fine-grained wetland classification for national wetland reserves using multi-source remote sensing data and Pixel Information Expert Engine (PIE-Engine) |
| title_short | Fine-grained wetland classification for national wetland reserves using multi-source remote sensing data and Pixel Information Expert Engine (PIE-Engine) |
| title_sort | fine grained wetland classification for national wetland reserves using multi source remote sensing data and pixel information expert engine pie engine |
| topic | Wetland mapping remote sensing ensemble learning pixel-based classification object-based classification PIE-Engine |
| url | https://www.tandfonline.com/doi/10.1080/15481603.2023.2286746 |
| work_keys_str_mv | AT hanliu finegrainedwetlandclassificationfornationalwetlandreservesusingmultisourceremotesensingdataandpixelinformationexpertenginepieengine AT tongkuiliao finegrainedwetlandclassificationfornationalwetlandreservesusingmultisourceremotesensingdataandpixelinformationexpertenginepieengine AT yuwang finegrainedwetlandclassificationfornationalwetlandreservesusingmultisourceremotesensingdataandpixelinformationexpertenginepieengine AT xiaomingqian finegrainedwetlandclassificationfornationalwetlandreservesusingmultisourceremotesensingdataandpixelinformationexpertenginepieengine AT xiaochenliu finegrainedwetlandclassificationfornationalwetlandreservesusingmultisourceremotesensingdataandpixelinformationexpertenginepieengine AT chengmingli finegrainedwetlandclassificationfornationalwetlandreservesusingmultisourceremotesensingdataandpixelinformationexpertenginepieengine AT shiweili finegrainedwetlandclassificationfornationalwetlandreservesusingmultisourceremotesensingdataandpixelinformationexpertenginepieengine AT zhanleiguan finegrainedwetlandclassificationfornationalwetlandreservesusingmultisourceremotesensingdataandpixelinformationexpertenginepieengine AT lijuezhu finegrainedwetlandclassificationfornationalwetlandreservesusingmultisourceremotesensingdataandpixelinformationexpertenginepieengine AT xiaoyuanzhou finegrainedwetlandclassificationfornationalwetlandreservesusingmultisourceremotesensingdataandpixelinformationexpertenginepieengine AT chongliu finegrainedwetlandclassificationfornationalwetlandreservesusingmultisourceremotesensingdataandpixelinformationexpertenginepieengine AT tengyunhu finegrainedwetlandclassificationfornationalwetlandreservesusingmultisourceremotesensingdataandpixelinformationexpertenginepieengine AT mingluo finegrainedwetlandclassificationfornationalwetlandreservesusingmultisourceremotesensingdataandpixelinformationexpertenginepieengine |