Motion estimation byusing stereo vision analysis for underwater observation system

Masyhuri Husna Binti Mazlan, Morisawa Daisuke, Koike Yoshikazu, Shimizu Junji, Enomoto Eriko, Hirohashi Noritaka, Shimizu Etsuro, Sakata Kunio

研究成果: Conference contribution

抄録

In this paper, motion estimation by using stereo vision analysis for underwater observation system (UOS) is presented. We developed a free-fall type UOS which was built with a glass sphere named Gyogyotto Camera. The UOS is employed to observe the underwater environment using the camera built inside of the glass sphere. The recorded images are used for stereo vision analysis to detect the motion estimation of the UOS. The proposed motion estimation uses Speed Up Robust Feature (SURF), stereo triangulation and Iterative Closest Point (ICP) algorithm. At first, calibration result with or without glass sphere on the land is described. The calibration result for underwater is also presented. Comparing the obtained results, we discussed the influence of the glass sphere. Moreover, we tried to estimate the motion of the camera by using the proposed method on the land and discuss the effective extraction of the feature points for underwater environment.

本文言語English
ホスト出版物のタイトルICEIC 2019 - International Conference on Electronics, Information, and Communication
出版社Institute of Electrical and Electronics Engineers Inc.
ISBN(電子版)9788995004449
DOI
出版ステータスPublished - 2019 5月 3
イベント18th International Conference on Electronics, Information, and Communication, ICEIC 2019 - Auckland, New Zealand
継続期間: 2019 1月 222019 1月 25

出版物シリーズ

名前ICEIC 2019 - International Conference on Electronics, Information, and Communication

Conference

Conference18th International Conference on Electronics, Information, and Communication, ICEIC 2019
国/地域New Zealand
CityAuckland
Period19/1/2219/1/25

ASJC Scopus subject areas

  • 制御およびシステム工学
  • 電子工学および電気工学

フィンガープリント

「Motion estimation byusing stereo vision analysis for underwater observation system」の研究トピックを掘り下げます。これらがまとまってユニークなフィンガープリントを構成します。

引用スタイル