2018 Fiscal Year Final Research Report
Wave height measurement using computer vision
| Project/Area Number |
16K18319
|
|---|
| Research Category |
Grant-in-Aid for Young Scientists (B)
|
| Allocation Type | Multi-year Fund |
|---|
| Research Field |
Naval and maritime engineering
|
|---|
| Research Institution | Tokyo University of Marine Science and Technology |
|---|
Principal Investigator |
|
|---|
| Project Period (FY) |
2016-04-01 – 2019-03-31
|
| Keywords | 画像処理 / 深層学習 |
|---|
| Outline of Final Research Achievements |
To measure the wave height using monocular camera, a system that the motion data and the front camera data of a ship are logged simultaneously is developed, and the data sets are collected for 30 hours. The front camera data is used as the input of deep neural network trained to estimate power spectrum of the ship motion.
In the beginning, the project had been planned to collect data in the area the wave height is measured by Japanese government, but the wave height meter was removed just after starting the project. Instead of actual wave height, now the motion spectrum of the ship is the prediction target of the neural network. Training neural network has not been succeeded yet, and the cause is investigated in detail.
|
| Free Research Field |
情報工学
|
| Academic Significance and Societal Importance of the Research Achievements |
船舶から用いる波高計には電波や音波を使ったものがあり高価である。これによらない場合、波高の計測は目視により行われている。コンピュータビジョンを用いて波高を計測することができれば波高計がより安価になる。ステレオカメラを用いた波高計測については多く研究されているが、本研究は単眼カメラでの波高計測を目指している。社会的・学術的意義としては、波高計が安価になり、客観的な波高計測手段が普及することで、より多くの場所で波高データが収集され、地球科学の発展に寄与すると思われる。
|
