| Project/Area Number |
23K13419
|
|---|
| Research Category |
Grant-in-Aid for Early-Career Scientists
|
| Allocation Type | Multi-year Fund |
|---|
| Review Section |
Basic Section 22050:Civil engineering plan and transportation engineering-related
|
|---|
| Research Institution | Tohoku University |
|---|
Principal Investigator |
袁 巍 東北大学, 災害科学国際研究所, 准教授 (60837475)
|
|---|
| Project Period (FY) |
2023-04-01 – 2026-03-31
|
| Project Status |
Granted (Fiscal Year 2023)
|
| Budget Amount *help |
¥4,550,000 (Direct Cost: ¥3,500,000、Indirect Cost: ¥1,050,000)
Fiscal Year 2025: ¥910,000 (Direct Cost: ¥700,000、Indirect Cost: ¥210,000)
Fiscal Year 2024: ¥1,690,000 (Direct Cost: ¥1,300,000、Indirect Cost: ¥390,000)
Fiscal Year 2023: ¥1,950,000 (Direct Cost: ¥1,500,000、Indirect Cost: ¥450,000)
|
|---|
| Keywords | Remote Sensing / Urban Mapping / Vision Transformer / DOM Generation / Depth Completion / Road Network Detection / Dense Image Matching / Graph Encoding / Photogrammetry / Change Detection / Data Fusion / Disaster assesment |
|---|
| Outline of Research at the Start |
Automatic post-disaster damage mapping is important, especially in Japan where natural disasters cause heavy losses frequently. Focusing on the perspective of physical damage mapping, this research proposed to achieve rapid, accurate, and cost-effective post-disaster damage mapping based on multi-modal remote sensing observations (and deep learning methods. There are three main steps in the proposed methods: 1) Multi-modal remote sensing data fusion for feature extraction; 2) Progressive-supervision for multi-task urban mapping; 3) transfer learning based post-disaster damage mapping.
|
| Outline of Annual Research Achievements |
This year, I initiated the research on Multi-task learning based post-disaster mapping via multi-modal remote sensing observations.I have achieved several publications: for the initial raw data processing,two papers on utilizing aerial and satellite image for DOM generation,mosaicking has been published on Geo-Spatial Information Science and Remote Sensing, and one paper utilizing deep diffusion model for few-shot depth information completion are published in CVPR 2023. For the urban scene mapping part,one paper on utilizing multi-constraint vision transformer for the urban building mapping has been pulished in IEEE J-STARS and two papers on deep neural network based urban infrastructure mapping are published in IGARSS 2023.
|
| Current Status of Research Progress |
Current Status of Research Progress
1: Research has progressed more than it was originally planned.
Reason
This year, I have moved smoothly on the developing the Multi-task learning based post-disaster mapping via multi-modal remote sensing observations. Although the complete post-disaster mapping is not finished yet, but most of the core components have been researched and initial methods such as DOM generation and urban mapping have been developed.The finished work has been published in prestigious conferences and journals.
|
| Strategy for Future Research Activity |
In the second year of this project, I planned to further conduct relevant research on the multi-modal remote sensing observation based urban mapping and 3D information generation as planned, and also try to have a better integration of these methods for post-disaster assesment such as change detection and pusedo earth observation data generation.
|
