| Project/Area Number |
22K14327
|
|---|
| Research Category |
Grant-in-Aid for Early-Career Scientists
|
| Allocation Type | Multi-year Fund |
|---|
| Review Section |
Basic Section 22040:Hydroengineering-related
|
|---|
| Research Institution | Tokyo University of Marine Science and Technology |
|---|
Principal Investigator |
呉 連慧 東京海洋大学, 学術研究院, 助教 (50907615)
|
|---|
| Project Period (FY) |
2022-04-01 – 2025-03-31
|
| Project Status |
Granted (Fiscal Year 2023)
|
| Budget Amount *help |
¥4,550,000 (Direct Cost: ¥3,500,000、Indirect Cost: ¥1,050,000)
Fiscal Year 2024: ¥650,000 (Direct Cost: ¥500,000、Indirect Cost: ¥150,000)
Fiscal Year 2023: ¥2,080,000 (Direct Cost: ¥1,600,000、Indirect Cost: ¥480,000)
Fiscal Year 2022: ¥1,820,000 (Direct Cost: ¥1,400,000、Indirect Cost: ¥420,000)
|
|---|
| Keywords | shoreline extraction / satellite SAR / sandy beach / coastal structure / X-band SAR / UAV / neural network / coast protection / Shoreline / Neural network |
|---|
| Outline of Research at the Start |
This research aims to develop a high-resolution shoreline extraction technique by using convolutional neural network (CNN) based on X-band SAR images. The developed technique is expected to improve our monitoring ability of the shoreline change for sustainable management of the coastal zone.
|
| Outline of Annual Research Achievements |
(1)The shoreline extraction model from Sentinel-1 SAR imagery based on DeepLab-v3+ was applied to 15 beaches in Japan to investigate its applicability and versatility. By using different combination of SAR images from beaches with varying characteristics, several models were constructed. The developed model demonstrates high accuracy in shoreline extraction, particularly when trained with a diverse set of images from various beaches.(2)A shoreline extraction model using 1-meter resolution X-band SAR images was proposed. Both for coasts with and without structures, the extraction accuray can be as high as 3-4 pixels. These results demonstrate the potential of using high-resolution X-band SAR images for coastal observations.
|
| Current Status of Research Progress |
Current Status of Research Progress
2: Research has progressed on the whole more than it was originally planned.
Reason
Shoreline extraction mdoel based on DeepLab-v3+ shows high accuary both for C-band and X-band SAR images. The applicability and versatility were throughoutly examined. Based on the above, it can be concluded that overall, progress is generally proceeding as planned.
|
| Strategy for Future Research Activity |
More X-band SAR images are expected to be collected to improve the model accuracy. In addition,the necessary amount of SAR data for construction of a shoreline extraction model for a specific sit will be investigated.
|
