| Project/Area Number |
22KF0301
|
|---|
| Project/Area Number (Other) |
22F21047 (2022)
|
|---|
| Research Category |
Grant-in-Aid for JSPS Fellows
|
| Allocation Type | Multi-year Fund (2023)
Single-year Grants (2022) |
|---|
| Section | 外国 |
|---|
| Review Section |
Basic Section 23030:Architectural planning and city planning-related
|
|---|
| Research Institution | Kyushu University |
|---|
Principal Investigator |
鶴崎 直樹 九州大学, 人間環境学研究院, 准教授 (20264096)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
MUHAMMAD MUHAMMAD 九州大学, 人間環境学研究院, 外国人特別研究員
|
|---|
| Project Period (FY) |
2023-03-08 – 2024-03-31
|
| Project Status |
Completed (Fiscal Year 2023)
|
| Budget Amount *help |
¥2,300,000 (Direct Cost: ¥2,300,000)
Fiscal Year 2023: ¥1,100,000 (Direct Cost: ¥1,100,000)
Fiscal Year 2022: ¥1,200,000 (Direct Cost: ¥1,200,000)
|
|---|
| Keywords | Land Use / Land Cover / deep learning / satellite imagery / Land Cover Change / Satellites Images / Machine Learning / Remote Sensing |
|---|
| Outline of Research at the Start |
本研究は、①最先端のディープラーニング技術とマルチスペクトル衛星画像を用いて“土地利用と土地被覆（LULC)”の変化のモニタリングと予測のための効果的なフレームワークを開発するとともに、②リモートセンシングデータからLULCの変化を検出し、気候変動や自然災害の影響を評価すること目指している。この成果は、正確な都市計画図や土地利用図等の情報インフラが整っていない国々における土地利用計画、資源
管理、災害管理など多分野での、より正確なLULCの変化のメカニズムを提供し、意思決定者が現状と予測される変化の傾向に応じた適切な土地利用政策の策定・実施に役立つものである。
|
| Outline of Annual Research Achievements |
In this study, we used Sentinel-2 satellite images (high-resolution multispectral data) from 2015 to 2021 taken by the European Space Agency (ESA) to detect land cover conditions and changes. We also utilized convolutional neural network (CNN) models based on deep learning (DL) techniques.
The objective of this study was to demonstrate the versatility and effectiveness of this method in addressing the challenges of land cover mapping across different contexts and landscapes in the earthquake affected areas of Mashiki Town, Kumamoto Prefecture and Greater Cairo, Egypt.
The innovative approach of combining deep learning and remote sensing techniques in this study proved to be very effective in tackling complex urban dynamics. A particularly striking achievement is the exceptional accuracy level achieved by the trained model. Across all land use and land cover classes in the study area, our model consistently achieved accuracy levels in excess of 90%.
This high level of accuracy demonstrates the robustness and reliability of our approach in accurately classifying and mapping land use and land cover change. We have utilized geospatial metrics and remote sensing methods to delve deeper into the dynamics of urban landscapes and provide a more comprehensive understanding of spatial change over time.
|
