Japan Geoscience Union Meeting 2022

Presentation information

[J] Oral

S (Solid Earth Sciences ) » S-TT Technology & Techniques

[S-TT39] Synthetic Aperture Radar and its application

Wed. May 25, 2022 10:45 AM - 12:15 PM 101 (International Conference Hall, Makuhari Messe)

convener:Takahiro Abe(Graduate School of Bioresources, Mie University ), convener:Yohei Kinoshita(University of Tsukuba), Yuji Himematsu(National Research Institute for Earth Science and Disaster Resilience), convener:Haemi Park(Japan Aerospace Exploration Agency), Chairperson:Yuji Himematsu(National Research Institute for Earth Science and Disaster Resilience), Haemi Park(Japan Aerospace Exploration Agency)

11:15 AM - 11:30 AM

[STT39-03] Landslides detection by machine learning and deep learning using Sentinel-1 intensity images

*Keisho Ito1, Yohei Kinoshita1 (1.Tsukuba University)


Keywords:Landslide, Synthetic Aperture Radar, Machine learning, Sentinel-1, Intensity image

When natural disasters such as earthquakes and heavy rains occur, multiple rapid landslides may occur. In addition to damaging to infrastructures and buildings, these landslides can cause isolation of the community and delay of rescue and evacuation, leading to serious damage in both human and economics. Therefore, it is necessary to detect disaster-affected areas quickly after the disaster occurrence. The satellite synthetic aperture radar (SAR) has the ability to obtain information over affected areas by observing the ground surface over wide area regardless of cloudy conditions. Recently, machine learning techniques has been applied to satellite images and are expected to be effective for the SAR data that is difficult to interpret by human's recognition. This study attempted to detect landslides by supervised machine classification including deep learning using Sentinel-1 intensity images.
We used multiple features such as backscatter intensity difference between the time-averaged images in the pre-disaster stage and a single intensity image that is available immediately after the disaster occurrence and terrain data for training. Similarly, we performed pixel-based machine learning using Gradient Boosting Decision Tree method and object-based deep learning using U-Net model, and then compared the results. U-Net is a semantic segmentation model, which can learn without losing the location information of objects compared to a simple FCN (fully convolution network). The result of landslide detection was performed for square areas with several hundred meters on a side, because there were a lot of small noises in the raw-resolution image and we thought the detailed shape and size of each landslides is not so important in practical use for administrations and municipalities. The training of the model and detection accuracy was evaluated on two cases, the 2018 Hokkaido Eastern Iburi Earthquake and the 2017 heavy rain event in Northern Kyushu.
In the results of machine learning using Gradient Boosting Decision Tree, the detection model could detect part of landslides, most of which were relatively large-size landslides in the case of 2018 Hokkaido Eastern Iburi Earthquake, achieving a Recall of 12.01%, a Precision of 58.87%, a Kappa coefficient of 0.20. On the other hand, there were many false landslide detections in 2017 heavy rain event in Northern Kyushu, possibly due to sudden surface change of scatter characteristics due to the strong rainfall. Similarly, we performed deep learning using the U-Net model and found that the accuracy was significantly improved compared to pixel-based machine learning, achieving a Kappa coefficient of 0.48. We confirmed an improvement in accuracy by deep learning in 2017 heavy rain event in Northern Kyushu, but there were still many incorrect detections compared to the earthquake landslide event in Iburi.