2020 Fiscal Year Annual Research Report
Unifying multiple RGB and depth cameras for real-time large-scale dynamic 3D modeling with unmanned micro aerial vehicles
| Project/Area Number |
19K20297
|
|---|
| Research Institution | Kyushu University |
|---|
Principal Investigator |
THOMAS DIEGO 九州大学, システム情報科学研究院, 助教 (10804651)
|
|---|
| Project Period (FY) |
2019-04-01 – 2021-03-31
|
| Keywords | RGB-D SLAM / Aerial drone / outdoor scene / 3D reconstruction / sensor fusion |
|---|
| Outline of Annual Research Achievements |
First year of the project
We proposed a new method for robust and accurate fusion of depth images. Our proposed framework achieved promising results for reconstructing accurate 3D models while using low computational power and being robust against misalignment errors without post-processing. Results published at 3DV 2019.
Depth information is not always available when reconstructing 3D models of outdoor scene. Therefore, at some time only 2D color images can be used. We proposed a solution to reconstruct 3D shape from 2D color images in the case of the human body. Our proposed network allows us to reconstruct detailed shapes of humans wearing loose clothes from single RGB images. Results published at CVPR 2020.
Last year of the project:
We propose a method for real-time human motion tracking and 3D body reconstruction that is cheap in memory consumption, handles fast motion and topological changes, while relatively simple to implement, and capable of producing 3D models with high accuracy. Results submitted for publication at ICIP 2021.
We proposed a new 3D scanning system equipped on a consumer-grade aerial drone that can capture live sequences of RGB-D data. our proposed system consists of a minicomputer powered by a portable battery and an RGB-D camera. We present a solution to easily equip this system onto a consumer-grade aerial drone (in a plug-and-play style), as well as the software to control it. We captured real world data and evaluated several state-of-the-art RGB-D SLAM techniques with our system. Results submitted for publication at IROS 2021 and MIRU 2021.
|
