2020 Fiscal Year Final Research Report
Development of a foundation for high-speed and highly-reliable automatic detection of occurrence of a natural disaster by using the satellite SAR data
| Project/Area Number |
19K22029
|
|---|
| Research Category |
Grant-in-Aid for Challenging Research (Exploratory)
|
| Allocation Type | Multi-year Fund |
|---|
| Review Section |
Medium-sized Section 25:Social systems engineering, safety engineering, disaster prevention engineering, and related fields
|
|---|
| Research Institution | Yamaguchi University |
|---|
Principal Investigator |
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
佐村 俊和 山口大学, 大学院創成科学研究科, 准教授 (30566617)
|
|---|
| Project Period (FY) |
2019-06-28 – 2021-03-31
|
| Keywords | リモートセンシング / 衛星SAR / 自然災害発災検知 / テータベース / 機械学習 |
|---|
| Outline of Final Research Achievements |
The purpose of this research theme is to develop a framework for automatic detection of natural disasters through satellite SAR data. We have got the following results through this research theme.
(1) We proposed a method for obtaining radar shadow area by using both the orbit of ALOS-2 and the 3D shape of earth's surface constructed from DEM data. (2) We proposed the concept of the small bounding box obtained from the hazard map on the DEM mesh as a unit of machine learning. (3) We proposed disaster damage detection methods from multitemporal SAR images by using convolutional neural networks.
|
| Free Research Field |
ビジュアルコンピューティング
|
| Academic Significance and Societal Importance of the Research Achievements |
レーダーシャドウ領域の明確化により発災判定対象からその領域を除外可能になり，衛星SARデータを利用した発災判定精度の向上が期待できる．また，従来人手により発災の有無を判定していたが，本研究課題推進により機械学習を利用した自動識別器によりある程度の信頼度で自動的に発災判定を行えることを示した．また，自動化の際に要求される処理の高速化の手段として発災判定小領域を単位とする方法を示すことができた．
|
