Automatic Mapping of Landslides by the ResU-Net
Massive landslides over large regions can be triggered by heavy rainfalls or major seismic events. Mapping regional landslides quickly is important for disaster mitigation. In recent years, deep learning methods have been successfully applied in many fields, including landslide automatic identificat...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2020-08-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/12/15/2487 |
| _version_ | 1797560491347804160 |
|---|---|
| author | Wenwen Qi Mengfei Wei Wentao Yang Chong Xu Chao Ma |
| author_facet | Wenwen Qi Mengfei Wei Wentao Yang Chong Xu Chao Ma |
| author_sort | Wenwen Qi |
| collection | DOAJ |
| description | Massive landslides over large regions can be triggered by heavy rainfalls or major seismic events. Mapping regional landslides quickly is important for disaster mitigation. In recent years, deep learning methods have been successfully applied in many fields, including landslide automatic identification. In this work, we proposed a deep learning approach, the ResU-Net, to map regional landslides automatically. This method and a baseline model (U-Net) were collectively tested in Tianshui city, Gansu province, where a heavy rainfall triggered more than 10,000 landslides in July 2013. All models were performed on a 3-band (near infrared, red, and green) GeoEye-1 image with a spatial resolution of 0.5 m. At such a fine spatial resolution, the study area is spatially heterogeneous. The tested study area is 128 km<sup>2</sup>, 80% of which was used to train models and the remaining 20% was used to validate accuracy of the models. This proposed ResU-Net achieved higher accuracy than the baseline U-Net model in this mountain region, where F1 improved by 0.09. Compared with the U-Net model, this proposed model (ResU-Net) performs better in discriminating landslides from bare floodplains along river valleys and unplanted terraces. By incorporating environmental information, this ResU-Net may also be applied to other landslide mapping, such as landslide susceptibility and hazard assessment. |
| first_indexed | 2024-03-10T18:00:20Z |
| format | Article |
| id | doaj.art-c1e01c9104494a1d8e75cb496ee47cf0 |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-10T18:00:20Z |
| publishDate | 2020-08-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-c1e01c9104494a1d8e75cb496ee47cf02023-11-20T08:55:42ZengMDPI AGRemote Sensing2072-42922020-08-011215248710.3390/rs12152487Automatic Mapping of Landslides by the ResU-NetWenwen Qi0Mengfei Wei1Wentao Yang2Chong Xu3Chao Ma4National Institute of Natural Hazards, Ministry of Emergency Management of China, Beijing 100085, ChinaTwenty-First Century Aerospace Technology Co., Ltd., Beijing 100096, ChinaSchool of Soil and Water Conservation, Beijing Forestry University, Beijing 100083, ChinaNational Institute of Natural Hazards, Ministry of Emergency Management of China, Beijing 100085, ChinaSchool of Soil and Water Conservation, Beijing Forestry University, Beijing 100083, ChinaMassive landslides over large regions can be triggered by heavy rainfalls or major seismic events. Mapping regional landslides quickly is important for disaster mitigation. In recent years, deep learning methods have been successfully applied in many fields, including landslide automatic identification. In this work, we proposed a deep learning approach, the ResU-Net, to map regional landslides automatically. This method and a baseline model (U-Net) were collectively tested in Tianshui city, Gansu province, where a heavy rainfall triggered more than 10,000 landslides in July 2013. All models were performed on a 3-band (near infrared, red, and green) GeoEye-1 image with a spatial resolution of 0.5 m. At such a fine spatial resolution, the study area is spatially heterogeneous. The tested study area is 128 km<sup>2</sup>, 80% of which was used to train models and the remaining 20% was used to validate accuracy of the models. This proposed ResU-Net achieved higher accuracy than the baseline U-Net model in this mountain region, where F1 improved by 0.09. Compared with the U-Net model, this proposed model (ResU-Net) performs better in discriminating landslides from bare floodplains along river valleys and unplanted terraces. By incorporating environmental information, this ResU-Net may also be applied to other landslide mapping, such as landslide susceptibility and hazard assessment.https://www.mdpi.com/2072-4292/12/15/2487regional landslide mappingremote sensingdeep learning models |
| spellingShingle | Wenwen Qi Mengfei Wei Wentao Yang Chong Xu Chao Ma Automatic Mapping of Landslides by the ResU-Net Remote Sensing regional landslide mapping remote sensing deep learning models |
| title | Automatic Mapping of Landslides by the ResU-Net |
| title_full | Automatic Mapping of Landslides by the ResU-Net |
| title_fullStr | Automatic Mapping of Landslides by the ResU-Net |
| title_full_unstemmed | Automatic Mapping of Landslides by the ResU-Net |
| title_short | Automatic Mapping of Landslides by the ResU-Net |
| title_sort | automatic mapping of landslides by the resu net |
| topic | regional landslide mapping remote sensing deep learning models |
| url | https://www.mdpi.com/2072-4292/12/15/2487 |
| work_keys_str_mv | AT wenwenqi automaticmappingoflandslidesbytheresunet AT mengfeiwei automaticmappingoflandslidesbytheresunet AT wentaoyang automaticmappingoflandslidesbytheresunet AT chongxu automaticmappingoflandslidesbytheresunet AT chaoma automaticmappingoflandslidesbytheresunet |