Distance Estimation Based on Vehicle Camera Images Using Deep Learning and Geometric Camera Model
深層学習とカメラ幾何を用いた車載カメラ画像での距離推定
- Delivery
- Available on this site
- Format
- Price
- Non-members (tax incl.):¥1,100 Members (tax incl.):¥880
- Publication code
- 20224262
- Paper/Info type
- Journal of Society of Automotive Engineers of Japan
Vol.76 No.4
- Pages
- 116-122(Total 7 p)
- Date of publication
- Apr 2022
- Publisher
- JSAE
- Language
- Japanese
Detailed Information
Category(J) | ホットトピックス Translation |
---|---|
Category(E) | Hot Topics |
Author(J) | 1) 佐々木 剛志, 2) 銭 智定, 3) 新 吉高, 4) 浅井 大輔 |
Author(E) | 1) Goshi Sasaki, 2) Trongmun Jiralerspong, 3) Yoshitaka Atarashi, 4) Daisuke Asai |
Affiliation(J) | 1) 日立製作所, 2) 日立製作所, 3) 日立製作所, 4) 日立製作所 |
Abstract(J) | 自動運転機能搭載車両の不具合解析には,走行シーンの把握が重要である.本稿では,シーン把握のために,低解像度の車載カメラ画像から前方対象物までの距離推定手法を提案する.提案手法では,超解像,物体検知,透視投影モデルを利用し,距離推定精度を維持しつつ,距離推定可能な対象の範囲を広げられることがわかった. Translation |
Abstract(E) | Understanding the surroundings in a driving scene is crucial in analyzing the defects of vehicle with autonomous functions. For the purpose of scene understanding, this article presents a method to estimate the distance between the leading and the ego vehicle from low resolution vehicle camera images. This method applies super-resolution, object detection and perspective projection techniques and is able to increase the number of target objects where their distance from the ego vehicle can be estimated while maintaining the estimation accuracy. |