Dynamic Detection of Forest Change in Hunan Province Based on Sentinel-2 Images and Deep Learning
Dynamic detection of forest change is the fundamental method of monitoring forest resources and an essential means of preserving the accuracy and timeliness of forest land resource data. This study focuses on a deep learning-based method for dynamic forest change detection using Sentinel-2 satellite...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-01-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/15/3/628 |
| _version_ | 1797623352704106496 |
|---|---|
| author | Jun Xiang Yuanjun Xing Wei Wei Enping Yan Jiawei Jiang Dengkui Mo |
| author_facet | Jun Xiang Yuanjun Xing Wei Wei Enping Yan Jiawei Jiang Dengkui Mo |
| author_sort | Jun Xiang |
| collection | DOAJ |
| description | Dynamic detection of forest change is the fundamental method of monitoring forest resources and an essential means of preserving the accuracy and timeliness of forest land resource data. This study focuses on a deep learning-based method for dynamic forest change detection using Sentinel-2 satellite data, especially within mountainous areas. First, the performance of various deep learning models (U-Net++, U-Net, LinkNet, DeepLabV3+, and STANet) and various loss functions (CrossEntropyLoss(CELoss), DiceLoss, FocalLoss, and their combinations) are compared on a self-made dataset. Next, the best model and loss function is used to predict the annual forest change in Hunan Province from 2017 to 2021, and the detection results are evaluated in 12 sample areas. Finally, forest changes are detected in Sentinel-2 images for each quarter of 2017–2021. In addition, a dynamic detection map of forest change in Hunan Province from 2017 to 2021 is drawn. The results reveal that the U-Net++ model and the CELoss performed the best on the self-made dataset, with a Precision of 0.795, a Recall of 0.748, and an <i>F</i>1-score of 0.771. The results of annual and quarterly forest change detection were consistent with the changes in the Sentinel-2 images with accurate boundaries. This result demonstrates the high practicality and generalizability of the method used in this paper. This paper achieves a rapid and accurate extraction of multi-temporal Sentinel-2 image forest change areas based on the U-Net++ model, which can be used as a benchmark for future large territorial areas monitoring and management of forest resources. |
| first_indexed | 2024-03-11T09:27:42Z |
| format | Article |
| id | doaj.art-bdc5fd3b3cf04451987c7514285f50ee |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-11T09:27:42Z |
| publishDate | 2023-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-bdc5fd3b3cf04451987c7514285f50ee2023-11-16T17:51:56ZengMDPI AGRemote Sensing2072-42922023-01-0115362810.3390/rs15030628Dynamic Detection of Forest Change in Hunan Province Based on Sentinel-2 Images and Deep LearningJun Xiang0Yuanjun Xing1Wei Wei2Enping Yan3Jiawei Jiang4Dengkui Mo5Key Laboratory of State Forestry and Grassland Administration on Forest Resources Management and Monitoring in Southern Area, Changsha 410004, ChinaCentral South Forest Inventory and Planning Institute of State Forestry Administration, Changsha 410004, ChinaForestry Research Institute of Guangxi Zhuang Autonomous Region, Nanning 530002, ChinaKey Laboratory of State Forestry and Grassland Administration on Forest Resources Management and Monitoring in Southern Area, Changsha 410004, ChinaKey Laboratory of State Forestry and Grassland Administration on Forest Resources Management and Monitoring in Southern Area, Changsha 410004, ChinaKey Laboratory of State Forestry and Grassland Administration on Forest Resources Management and Monitoring in Southern Area, Changsha 410004, ChinaDynamic detection of forest change is the fundamental method of monitoring forest resources and an essential means of preserving the accuracy and timeliness of forest land resource data. This study focuses on a deep learning-based method for dynamic forest change detection using Sentinel-2 satellite data, especially within mountainous areas. First, the performance of various deep learning models (U-Net++, U-Net, LinkNet, DeepLabV3+, and STANet) and various loss functions (CrossEntropyLoss(CELoss), DiceLoss, FocalLoss, and their combinations) are compared on a self-made dataset. Next, the best model and loss function is used to predict the annual forest change in Hunan Province from 2017 to 2021, and the detection results are evaluated in 12 sample areas. Finally, forest changes are detected in Sentinel-2 images for each quarter of 2017–2021. In addition, a dynamic detection map of forest change in Hunan Province from 2017 to 2021 is drawn. The results reveal that the U-Net++ model and the CELoss performed the best on the self-made dataset, with a Precision of 0.795, a Recall of 0.748, and an <i>F</i>1-score of 0.771. The results of annual and quarterly forest change detection were consistent with the changes in the Sentinel-2 images with accurate boundaries. This result demonstrates the high practicality and generalizability of the method used in this paper. This paper achieves a rapid and accurate extraction of multi-temporal Sentinel-2 image forest change areas based on the U-Net++ model, which can be used as a benchmark for future large territorial areas monitoring and management of forest resources.https://www.mdpi.com/2072-4292/15/3/628dynamic detectionforest changedeep learningSentinel-2Hunan Province |
| spellingShingle | Jun Xiang Yuanjun Xing Wei Wei Enping Yan Jiawei Jiang Dengkui Mo Dynamic Detection of Forest Change in Hunan Province Based on Sentinel-2 Images and Deep Learning Remote Sensing dynamic detection forest change deep learning Sentinel-2 Hunan Province |
| title | Dynamic Detection of Forest Change in Hunan Province Based on Sentinel-2 Images and Deep Learning |
| title_full | Dynamic Detection of Forest Change in Hunan Province Based on Sentinel-2 Images and Deep Learning |
| title_fullStr | Dynamic Detection of Forest Change in Hunan Province Based on Sentinel-2 Images and Deep Learning |
| title_full_unstemmed | Dynamic Detection of Forest Change in Hunan Province Based on Sentinel-2 Images and Deep Learning |
| title_short | Dynamic Detection of Forest Change in Hunan Province Based on Sentinel-2 Images and Deep Learning |
| title_sort | dynamic detection of forest change in hunan province based on sentinel 2 images and deep learning |
| topic | dynamic detection forest change deep learning Sentinel-2 Hunan Province |
| url | https://www.mdpi.com/2072-4292/15/3/628 |
| work_keys_str_mv | AT junxiang dynamicdetectionofforestchangeinhunanprovincebasedonsentinel2imagesanddeeplearning AT yuanjunxing dynamicdetectionofforestchangeinhunanprovincebasedonsentinel2imagesanddeeplearning AT weiwei dynamicdetectionofforestchangeinhunanprovincebasedonsentinel2imagesanddeeplearning AT enpingyan dynamicdetectionofforestchangeinhunanprovincebasedonsentinel2imagesanddeeplearning AT jiaweijiang dynamicdetectionofforestchangeinhunanprovincebasedonsentinel2imagesanddeeplearning AT dengkuimo dynamicdetectionofforestchangeinhunanprovincebasedonsentinel2imagesanddeeplearning |