2:00 PM - 2:15 PM
[STT39-07] Flood area Extraction around the Tokigawa-river at the Typhoon No. 19 event in 2019 using SAR
Keywords:SAR, flood area
Typhoon No. 19 in the first year of Reiwa caused record heavy rains in Shizuoka prefecture, Kanto region, Koshinetsu region, Tohoku region, etc. It was the largest damage ever. In recent years, the probability of heavy rainfall has increased due to global warming, and many heavy rainfall disasters have occurred. It is predicted that the frequency of heavy rainfall and short-term heavy rainfall and the amount of precipitation will increase in the future, and there is growing concern about the occurrence of large-scale floods. On the other hand, Flood damage surveys, on the other hand, entail a large number of personnel and great danger. Therefore, the theme of this research was to extract the water surface reflected by satellite images by taking advantage of the two merits of the all-weather type and the synthetic aperture radar (SAR) that enables wide-area observation. As a procedure,
(1) From the SAR image taken at the time of the disaster of typhoon No. 19, the backscattering coefficient (σ0) of the point assumed to be inundated area and the backscattering coefficient (σ0) of the point assumed to be non-inundated area are obtained. Measure each. Next, the point where the distribution of the backscattering coefficient in the inundated area and the distribution of the backscattering coefficient in the non-inundated area intersect is set as the threshold value (σth), and the classification algorithm of the water area and the land area by the threshold method is created.
(2) Based on the classification algorithm, SAR images observed by the European Space Agency (ESA) Sentinel-1 satellite (VH, VV polarization) and the Japan Aerospace Exploration Agency (JAXA) ALOS-2 (HH polarization) in 2021 It is applied, the accuracy of water area detection is scrutinized, and the validity of the threshold method method and the validity of the threshold itself are verified.
(3) Apply the created classification algorithm at the time of Typhoon No. 19 and examine it.
I went in the flow.
As a result, the Sentinel-1 satellite (VH polarization) showed the highest overall accuracy for land and water areas compared to the verification using the NDWI, with a threshold of -19.93 dB and an overall accuracy of 97.70 %. The satellite (VV polarization) recorded a total accuracy of 97.52 % at a threshold of -14.04 dB, and the ALOS-2 satellite (HH polarization) recorded a total accuracy of 98.11 % at a threshold of -8.92 dB. In addition, when these thresholds were applied to SAR images at the time of Typhoon No. 19 in 2019 and verified, an accuracy of 66.79 % was recorded. This difference in accuracy is considered to be due to a time lag of approximately 34 hours between the survey date and time of the inundation estimation stage map adopted by the Geographical Survey Institute as the true value and the observation date and time by SAR.
From this study, it was found that it is possible to detect the open water surface using the threshold method.
As future developments, we will work on changes in the amount of flooding and flooded areas in the gently sloping terrain that is unique to overseas, improving the accuracy of dividing land and water areas, and improving the accuracy of detecting flooded areas caused by flood damage.
(1) From the SAR image taken at the time of the disaster of typhoon No. 19, the backscattering coefficient (σ0) of the point assumed to be inundated area and the backscattering coefficient (σ0) of the point assumed to be non-inundated area are obtained. Measure each. Next, the point where the distribution of the backscattering coefficient in the inundated area and the distribution of the backscattering coefficient in the non-inundated area intersect is set as the threshold value (σth), and the classification algorithm of the water area and the land area by the threshold method is created.
(2) Based on the classification algorithm, SAR images observed by the European Space Agency (ESA) Sentinel-1 satellite (VH, VV polarization) and the Japan Aerospace Exploration Agency (JAXA) ALOS-2 (HH polarization) in 2021 It is applied, the accuracy of water area detection is scrutinized, and the validity of the threshold method method and the validity of the threshold itself are verified.
(3) Apply the created classification algorithm at the time of Typhoon No. 19 and examine it.
I went in the flow.
As a result, the Sentinel-1 satellite (VH polarization) showed the highest overall accuracy for land and water areas compared to the verification using the NDWI, with a threshold of -19.93 dB and an overall accuracy of 97.70 %. The satellite (VV polarization) recorded a total accuracy of 97.52 % at a threshold of -14.04 dB, and the ALOS-2 satellite (HH polarization) recorded a total accuracy of 98.11 % at a threshold of -8.92 dB. In addition, when these thresholds were applied to SAR images at the time of Typhoon No. 19 in 2019 and verified, an accuracy of 66.79 % was recorded. This difference in accuracy is considered to be due to a time lag of approximately 34 hours between the survey date and time of the inundation estimation stage map adopted by the Geographical Survey Institute as the true value and the observation date and time by SAR.
From this study, it was found that it is possible to detect the open water surface using the threshold method.
As future developments, we will work on changes in the amount of flooding and flooded areas in the gently sloping terrain that is unique to overseas, improving the accuracy of dividing land and water areas, and improving the accuracy of detecting flooded areas caused by flood damage.