Traffic Light Detection for Driving Environment Recognition Using Omnidirectional Camera
全方位カメラを用いた走行環境認識のための信号機検出に関する研究
- Delivery
- Available on the other site
- Click here to order.
- Publication code
- 20214264
- Paper/Info type
- JSAE Transaction
Vol.52 No.3
- Pages
- 695-700(Total 6 p)
- Date of publication
- May 2021
- Publisher
- JSAE
- Language
- Japanese
Detailed Information
| Category(J) | 研究論文 Translation |
|---|---|
| Category(E) | ResearchPaper |
| Author(J) | 1) 石田 泰之, 2) Mohd Hafiz Hilman bin Mohammad Sofian , 3) 伊東 敏夫 |
| Author(E) | 1) Yasuyuki Ishida, 2) Mohd Hafiz Hilman bin Mohammad Sofian, 3) Toshio Ito |
| Affiliation(J) | 1) 芝浦工業大学, 2) 芝浦工業大学, 3) 芝浦工業大学 |
| Abstract(J) | 普及型の自動運転車では，最小限のセンサ構成が要求される．そのため，広い視野を持つ全方位カメラが適している．しかし，全方位カメラから得られた画像は歪むため，既存の畳み込みニューラルネットワークの適用が困難である．そこで，本研究では全方位カメラの歪みに対応した信号機検出手法を提案する． Translation |
| Abstract(E) | Object Detection is an important part of autonomous driving system. A popular type of autonomous vehicle requires a minimum sensor configuration. Since omnidirectional camera has a large angle of view, it is suitable for autonomous vehicle. However, the omnidirectional image is distorted in the shape. Therefore, it is difficult to apply existing image processing and machine learning. In this research, we focus on the fact that the omnidirectional image is distorted in the shape but the color does not change and we use neural networks for object detection, considering that neural networks can learn distorted shapes. |
