Estimation of Tropical Cyclone Intensity Using Multi-Platform Remote Sensing and Deep Learning with Environmental Field Information
Accurate tropical cyclone (TC) intensity estimation is crucial for prediction and disaster prevention. Currently, significant progress has been achieved for the application of convolutional neural networks (CNNs) in TC intensity estimation. However, many studies have overlooked the fact that the loc...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-04-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/15/8/2085 |
| _version_ | 1797603590039142400 |
|---|---|
| author | Wei Tian Linhong Lai Xianghua Niu Xinxin Zhou Yonghong Zhang Lim Kam Sian Thiam Choy Kenny |
| author_facet | Wei Tian Linhong Lai Xianghua Niu Xinxin Zhou Yonghong Zhang Lim Kam Sian Thiam Choy Kenny |
| author_sort | Wei Tian |
| collection | DOAJ |
| description | Accurate tropical cyclone (TC) intensity estimation is crucial for prediction and disaster prevention. Currently, significant progress has been achieved for the application of convolutional neural networks (CNNs) in TC intensity estimation. However, many studies have overlooked the fact that the local convolution used by CNNs does not consider the global spatial relationships between pixels. Hence, they can only capture limited spatial contextual information. In addition, the special rotation invariance and symmetry structure of TC cannot be fully expressed by convolutional kernels alone. Therefore, this study proposes a new deep learning-based model for TC intensity estimation, which uses a combination of rotation equivariant convolution and Transformer to address the rotation invariance and symmetry structure of TC. Combining the two can allow capturing both local and global spatial contextual information, thereby achieving more accurate intensity estimation. Furthermore, we fused multi-platform satellite remote sensing data into the model to provide more information about the TC structure. At the same time, we integrate the physical environmental field information into the model, which can help capture the impact of these external factors on TC intensity and further improve the estimation accuracy. Finally, we use TCs from 2003 to 2015 to train our model and use 2016 and 2017 data as independent validation sets to verify our model. The overall root mean square error (RMSE) is 8.19 kt. For a subset of 482 samples (from the East Pacific and Atlantic) observed by aircraft reconnaissance, the root mean square error is 7.88 kt. |
| first_indexed | 2024-03-11T04:34:13Z |
| format | Article |
| id | doaj.art-9d3bc49f841c40a09dc03d73a1672ac1 |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-11T04:34:13Z |
| publishDate | 2023-04-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-9d3bc49f841c40a09dc03d73a1672ac12023-11-17T21:11:50ZengMDPI AGRemote Sensing2072-42922023-04-01158208510.3390/rs15082085Estimation of Tropical Cyclone Intensity Using Multi-Platform Remote Sensing and Deep Learning with Environmental Field InformationWei Tian0Linhong Lai1Xianghua Niu2Xinxin Zhou3Yonghong Zhang4Lim Kam Sian Thiam Choy Kenny5School of Computer and Softwar, Nanjing University of Information Science and Technology, No. 219, Ningliu Road, Nanjing 210044, ChinaSchool of Computer and Softwar, Nanjing University of Information Science and Technology, No. 219, Ningliu Road, Nanjing 210044, ChinaState Key Laboratory of Geo-Information Engineering, Xi’an 710054, ChinaSchool of Computer and Softwar, Nanjing University of Information Science and Technology, No. 219, Ningliu Road, Nanjing 210044, ChinaSchool of Automation, Nanjing University of Information Science and Technology, No. 219, Ningliu Road, Nanjing 210044, ChinaSchool of Atmospheric Science and Remote Sensing, Wuxi University, 333 Xishan Avenue, Wuxi 214105, ChinaAccurate tropical cyclone (TC) intensity estimation is crucial for prediction and disaster prevention. Currently, significant progress has been achieved for the application of convolutional neural networks (CNNs) in TC intensity estimation. However, many studies have overlooked the fact that the local convolution used by CNNs does not consider the global spatial relationships between pixels. Hence, they can only capture limited spatial contextual information. In addition, the special rotation invariance and symmetry structure of TC cannot be fully expressed by convolutional kernels alone. Therefore, this study proposes a new deep learning-based model for TC intensity estimation, which uses a combination of rotation equivariant convolution and Transformer to address the rotation invariance and symmetry structure of TC. Combining the two can allow capturing both local and global spatial contextual information, thereby achieving more accurate intensity estimation. Furthermore, we fused multi-platform satellite remote sensing data into the model to provide more information about the TC structure. At the same time, we integrate the physical environmental field information into the model, which can help capture the impact of these external factors on TC intensity and further improve the estimation accuracy. Finally, we use TCs from 2003 to 2015 to train our model and use 2016 and 2017 data as independent validation sets to verify our model. The overall root mean square error (RMSE) is 8.19 kt. For a subset of 482 samples (from the East Pacific and Atlantic) observed by aircraft reconnaissance, the root mean square error is 7.88 kt.https://www.mdpi.com/2072-4292/15/8/2085tropical cyclone intensitymulti-platform remote sensing data fusionremote sensingrotation equivariant convolutionattention mechanism and transformer |
| spellingShingle | Wei Tian Linhong Lai Xianghua Niu Xinxin Zhou Yonghong Zhang Lim Kam Sian Thiam Choy Kenny Estimation of Tropical Cyclone Intensity Using Multi-Platform Remote Sensing and Deep Learning with Environmental Field Information Remote Sensing tropical cyclone intensity multi-platform remote sensing data fusion remote sensing rotation equivariant convolution attention mechanism and transformer |
| title | Estimation of Tropical Cyclone Intensity Using Multi-Platform Remote Sensing and Deep Learning with Environmental Field Information |
| title_full | Estimation of Tropical Cyclone Intensity Using Multi-Platform Remote Sensing and Deep Learning with Environmental Field Information |
| title_fullStr | Estimation of Tropical Cyclone Intensity Using Multi-Platform Remote Sensing and Deep Learning with Environmental Field Information |
| title_full_unstemmed | Estimation of Tropical Cyclone Intensity Using Multi-Platform Remote Sensing and Deep Learning with Environmental Field Information |
| title_short | Estimation of Tropical Cyclone Intensity Using Multi-Platform Remote Sensing and Deep Learning with Environmental Field Information |
| title_sort | estimation of tropical cyclone intensity using multi platform remote sensing and deep learning with environmental field information |
| topic | tropical cyclone intensity multi-platform remote sensing data fusion remote sensing rotation equivariant convolution attention mechanism and transformer |
| url | https://www.mdpi.com/2072-4292/15/8/2085 |
| work_keys_str_mv | AT weitian estimationoftropicalcycloneintensityusingmultiplatformremotesensinganddeeplearningwithenvironmentalfieldinformation AT linhonglai estimationoftropicalcycloneintensityusingmultiplatformremotesensinganddeeplearningwithenvironmentalfieldinformation AT xianghuaniu estimationoftropicalcycloneintensityusingmultiplatformremotesensinganddeeplearningwithenvironmentalfieldinformation AT xinxinzhou estimationoftropicalcycloneintensityusingmultiplatformremotesensinganddeeplearningwithenvironmentalfieldinformation AT yonghongzhang estimationoftropicalcycloneintensityusingmultiplatformremotesensinganddeeplearningwithenvironmentalfieldinformation AT limkamsianthiamchoykenny estimationoftropicalcycloneintensityusingmultiplatformremotesensinganddeeplearningwithenvironmentalfieldinformation |