Self-supervised deep learning for detection of forest disturbance types in a subtropical ecosystem using transformer and Sentinel-1 and Sentinel-2 time series data
Accurate and timely detection of forest disturbance types is crucial for evaluating ecosystem health and global climate stability. Time series remote-sensing data offers valuable spatiotemporal information. However, frequent cloud cover in subtropical regions disrupts the temporal consistency of opt...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Taylor & Francis Group
2025-12-01
|
| Series: | International Journal of Digital Earth |
| Subjects: | |
| Online Access: | https://www.tandfonline.com/doi/10.1080/17538947.2025.2537325 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1839619808574832640 |
|---|---|
| author | Ming Zhang Guiying Li Dengsheng Lu Cong Xu Haotian Zhao Dengqiu Li |
| author_facet | Ming Zhang Guiying Li Dengsheng Lu Cong Xu Haotian Zhao Dengqiu Li |
| author_sort | Ming Zhang |
| collection | DOAJ |
| description | Accurate and timely detection of forest disturbance types is crucial for evaluating ecosystem health and global climate stability. Time series remote-sensing data offers valuable spatiotemporal information. However, frequent cloud cover in subtropical regions disrupts the temporal consistency of optical satellite data. In addition, the impact of different data sources on modeling accuracy and the challenge of acquiring large and labeled datasets for deep learning are considerable obstacles. In this study, a novel positional encoding module was designed to handle the irregular Sentinel-2 time series. The Transformer model based on this positional encoding module effectively fused Sentinel-1 and Sentinel-2 data. Meanwhile, self-supervised learning was used to address the issue of insufficient samples. The improved Transformer model was successfully applied to detect clear-cutting and disease disturbances with a F1 score of 0.95. Our results highlighted that appropriate encoding techniques increased the model’s performance by between 5% and 14%. This research also found that while Sentinel-2 data alone yields good accuracy, combining Sentinel-1 and Sentinel-2 data improved model accuracy. The self-supervised learning model achieved higher accuracy (0.95) than the supervised model (0.86) on a limited data (20% of the training samples) and exhibited stable accuracy across different training data proportions. |
| format | Article |
| id | doaj-art-14ca25c49d1640208d49c9ad6d35f7d0 |
| institution | Matheson Library |
| issn | 1753-8947 1753-8955 |
| language | English |
| publishDate | 2025-12-01 |
| publisher | Taylor & Francis Group |
| record_format | Article |
| series | International Journal of Digital Earth |
| spelling | doaj-art-14ca25c49d1640208d49c9ad6d35f7d02025-07-23T00:55:57ZengTaylor & Francis GroupInternational Journal of Digital Earth1753-89471753-89552025-12-0118110.1080/17538947.2025.2537325Self-supervised deep learning for detection of forest disturbance types in a subtropical ecosystem using transformer and Sentinel-1 and Sentinel-2 time series dataMing Zhang0Guiying Li1Dengsheng Lu2Cong Xu3Haotian Zhao4Dengqiu Li5Fujian Provincial Key Laboratory for Subtropical Resources and Environment, Fujian Normal University, Fuzhou, People’s Republic of ChinaFujian Provincial Key Laboratory for Subtropical Resources and Environment, Fujian Normal University, Fuzhou, People’s Republic of ChinaFujian Provincial Key Laboratory for Subtropical Resources and Environment, Fujian Normal University, Fuzhou, People’s Republic of ChinaSchool of Forestry, University of Canterbury, Christchurch, New ZealandSchool of Forestry, University of Canterbury, Christchurch, New ZealandFujian Provincial Key Laboratory for Subtropical Resources and Environment, Fujian Normal University, Fuzhou, People’s Republic of ChinaAccurate and timely detection of forest disturbance types is crucial for evaluating ecosystem health and global climate stability. Time series remote-sensing data offers valuable spatiotemporal information. However, frequent cloud cover in subtropical regions disrupts the temporal consistency of optical satellite data. In addition, the impact of different data sources on modeling accuracy and the challenge of acquiring large and labeled datasets for deep learning are considerable obstacles. In this study, a novel positional encoding module was designed to handle the irregular Sentinel-2 time series. The Transformer model based on this positional encoding module effectively fused Sentinel-1 and Sentinel-2 data. Meanwhile, self-supervised learning was used to address the issue of insufficient samples. The improved Transformer model was successfully applied to detect clear-cutting and disease disturbances with a F1 score of 0.95. Our results highlighted that appropriate encoding techniques increased the model’s performance by between 5% and 14%. This research also found that while Sentinel-2 data alone yields good accuracy, combining Sentinel-1 and Sentinel-2 data improved model accuracy. The self-supervised learning model achieved higher accuracy (0.95) than the supervised model (0.86) on a limited data (20% of the training samples) and exhibited stable accuracy across different training data proportions.https://www.tandfonline.com/doi/10.1080/17538947.2025.2537325Forest disturbance typestime seriesSentinel-1/2self-supervised deep learningTransformer |
| spellingShingle | Ming Zhang Guiying Li Dengsheng Lu Cong Xu Haotian Zhao Dengqiu Li Self-supervised deep learning for detection of forest disturbance types in a subtropical ecosystem using transformer and Sentinel-1 and Sentinel-2 time series data International Journal of Digital Earth Forest disturbance types time series Sentinel-1/2 self-supervised deep learning Transformer |
| title | Self-supervised deep learning for detection of forest disturbance types in a subtropical ecosystem using transformer and Sentinel-1 and Sentinel-2 time series data |
| title_full | Self-supervised deep learning for detection of forest disturbance types in a subtropical ecosystem using transformer and Sentinel-1 and Sentinel-2 time series data |
| title_fullStr | Self-supervised deep learning for detection of forest disturbance types in a subtropical ecosystem using transformer and Sentinel-1 and Sentinel-2 time series data |
| title_full_unstemmed | Self-supervised deep learning for detection of forest disturbance types in a subtropical ecosystem using transformer and Sentinel-1 and Sentinel-2 time series data |
| title_short | Self-supervised deep learning for detection of forest disturbance types in a subtropical ecosystem using transformer and Sentinel-1 and Sentinel-2 time series data |
| title_sort | self supervised deep learning for detection of forest disturbance types in a subtropical ecosystem using transformer and sentinel 1 and sentinel 2 time series data |
| topic | Forest disturbance types time series Sentinel-1/2 self-supervised deep learning Transformer |
| url | https://www.tandfonline.com/doi/10.1080/17538947.2025.2537325 |
| work_keys_str_mv | AT mingzhang selfsuperviseddeeplearningfordetectionofforestdisturbancetypesinasubtropicalecosystemusingtransformerandsentinel1andsentinel2timeseriesdata AT guiyingli selfsuperviseddeeplearningfordetectionofforestdisturbancetypesinasubtropicalecosystemusingtransformerandsentinel1andsentinel2timeseriesdata AT dengshenglu selfsuperviseddeeplearningfordetectionofforestdisturbancetypesinasubtropicalecosystemusingtransformerandsentinel1andsentinel2timeseriesdata AT congxu selfsuperviseddeeplearningfordetectionofforestdisturbancetypesinasubtropicalecosystemusingtransformerandsentinel1andsentinel2timeseriesdata AT haotianzhao selfsuperviseddeeplearningfordetectionofforestdisturbancetypesinasubtropicalecosystemusingtransformerandsentinel1andsentinel2timeseriesdata AT dengqiuli selfsuperviseddeeplearningfordetectionofforestdisturbancetypesinasubtropicalecosystemusingtransformerandsentinel1andsentinel2timeseriesdata |