Forest Fire Segmentation from Aerial Imagery Data Using an Improved Instance Segmentation Model
In recent years, forest-fire monitoring methods represented by deep learning have been developed rapidly. The use of drone technology and optimization of existing models to improve forest-fire recognition accuracy and segmentation quality are of great significance for understanding the spatial distr...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-07-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/14/13/3159 |
| _version_ | 1797434016900579328 |
|---|---|
| author | Zhihao Guan Xinyu Miao Yunjie Mu Quan Sun Qiaolin Ye Demin Gao |
| author_facet | Zhihao Guan Xinyu Miao Yunjie Mu Quan Sun Qiaolin Ye Demin Gao |
| author_sort | Zhihao Guan |
| collection | DOAJ |
| description | In recent years, forest-fire monitoring methods represented by deep learning have been developed rapidly. The use of drone technology and optimization of existing models to improve forest-fire recognition accuracy and segmentation quality are of great significance for understanding the spatial distribution of forest fires and protecting forest resources. Due to the spreading and irregular nature of fire, it is extremely tough to detect fire accurately in a complex environment. Based on the aerial imagery dataset FLAME, this paper focuses on the analysis of methods to two deep-learning problems: (1) the video frames are classified as two classes (fire, no-fire) according to the presence or absence of fire. A novel image classification method based on channel domain attention mechanism was developed, which achieved a classification accuracy of 93.65%. (2) We propose a novel instance segmentation method (MaskSU R-CNN) for incipient forest-fire detection and segmentation based on MS R-CNN model. For the optimized model, the MaskIoU branch is reconstructed by a U-shaped network in order to reduce the segmentation error. Experimental results show that the precision of our MaskSU R-CNN reached 91.85%, recall 88.81%, F1-score 90.30%, and mean intersection over union (<i>mIoU</i>) 82.31%. Compared with many state-of-the-art segmentation models, our method achieves satisfactory results on forest-fire dataset. |
| first_indexed | 2024-03-09T10:25:17Z |
| format | Article |
| id | doaj.art-7686ce7e20b74b5e842f8e8880d0ef98 |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-09T10:25:17Z |
| publishDate | 2022-07-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-7686ce7e20b74b5e842f8e8880d0ef982023-12-01T21:40:58ZengMDPI AGRemote Sensing2072-42922022-07-011413315910.3390/rs14133159Forest Fire Segmentation from Aerial Imagery Data Using an Improved Instance Segmentation ModelZhihao Guan0Xinyu Miao1Yunjie Mu2Quan Sun3Qiaolin Ye4Demin Gao5College of Information Science and Technology, Nanjing Forestry University, Nanjing 210037, ChinaCollege of Information Science and Technology, Nanjing Forestry University, Nanjing 210037, ChinaCollege of Information Science and Technology, Nanjing Forestry University, Nanjing 210037, ChinaCollege of Information Science and Technology, Nanjing Forestry University, Nanjing 210037, ChinaCollege of Information Science and Technology, Nanjing Forestry University, Nanjing 210037, ChinaCollege of Information Science and Technology, Nanjing Forestry University, Nanjing 210037, ChinaIn recent years, forest-fire monitoring methods represented by deep learning have been developed rapidly. The use of drone technology and optimization of existing models to improve forest-fire recognition accuracy and segmentation quality are of great significance for understanding the spatial distribution of forest fires and protecting forest resources. Due to the spreading and irregular nature of fire, it is extremely tough to detect fire accurately in a complex environment. Based on the aerial imagery dataset FLAME, this paper focuses on the analysis of methods to two deep-learning problems: (1) the video frames are classified as two classes (fire, no-fire) according to the presence or absence of fire. A novel image classification method based on channel domain attention mechanism was developed, which achieved a classification accuracy of 93.65%. (2) We propose a novel instance segmentation method (MaskSU R-CNN) for incipient forest-fire detection and segmentation based on MS R-CNN model. For the optimized model, the MaskIoU branch is reconstructed by a U-shaped network in order to reduce the segmentation error. Experimental results show that the precision of our MaskSU R-CNN reached 91.85%, recall 88.81%, F1-score 90.30%, and mean intersection over union (<i>mIoU</i>) 82.31%. Compared with many state-of-the-art segmentation models, our method achieves satisfactory results on forest-fire dataset.https://www.mdpi.com/2072-4292/14/13/3159fire recognitioninstance segmentationcomputer visiondeep learningaerial imagery |
| spellingShingle | Zhihao Guan Xinyu Miao Yunjie Mu Quan Sun Qiaolin Ye Demin Gao Forest Fire Segmentation from Aerial Imagery Data Using an Improved Instance Segmentation Model Remote Sensing fire recognition instance segmentation computer vision deep learning aerial imagery |
| title | Forest Fire Segmentation from Aerial Imagery Data Using an Improved Instance Segmentation Model |
| title_full | Forest Fire Segmentation from Aerial Imagery Data Using an Improved Instance Segmentation Model |
| title_fullStr | Forest Fire Segmentation from Aerial Imagery Data Using an Improved Instance Segmentation Model |
| title_full_unstemmed | Forest Fire Segmentation from Aerial Imagery Data Using an Improved Instance Segmentation Model |
| title_short | Forest Fire Segmentation from Aerial Imagery Data Using an Improved Instance Segmentation Model |
| title_sort | forest fire segmentation from aerial imagery data using an improved instance segmentation model |
| topic | fire recognition instance segmentation computer vision deep learning aerial imagery |
| url | https://www.mdpi.com/2072-4292/14/13/3159 |
| work_keys_str_mv | AT zhihaoguan forestfiresegmentationfromaerialimagerydatausinganimprovedinstancesegmentationmodel AT xinyumiao forestfiresegmentationfromaerialimagerydatausinganimprovedinstancesegmentationmodel AT yunjiemu forestfiresegmentationfromaerialimagerydatausinganimprovedinstancesegmentationmodel AT quansun forestfiresegmentationfromaerialimagerydatausinganimprovedinstancesegmentationmodel AT qiaolinye forestfiresegmentationfromaerialimagerydatausinganimprovedinstancesegmentationmodel AT demingao forestfiresegmentationfromaerialimagerydatausinganimprovedinstancesegmentationmodel |