Proposal of Acquisition Method for Resource Information of Boxing Marine Products in Offshore Trawl Fishery
-
- Tomoo Shiigi
- National Fisheries University
-
- Kazuhiro Tokunaga
- National Fisheries University
-
- Hirofumi Matsumoto
- National Fisheries University
-
- Yohei Miyamoto
- Showa Suisan Co. Ltd.
抄録
In an offshore trawl fishery, the use of ICT (Information and Communication Technology) for a resource management has been important issue, because it is required that a lot of effort to grasp the catch of a large number and various species of marine products. In this study, we propose a method of auto-estimating information such as fish species and size from packed fish in the box using machine vision and information processing technology as a system to support the resource management in the offshore trawl fisheries. The proposed method uses a convolutional neural network to discriminate fish species from color images of packed fish in the box. On the other hand, the number and size of fish are estimated from the depth image using signal processing. In this presentation, we explain the method and show the experimental results.
収録刊行物
-
- IEICE Proceeding Series
-
IEICE Proceeding Series 66 77-82, 2021-09-20
The Institute of Electronics, Information and Communication Engineers
- Tweet
詳細情報 詳細情報について
-
- CRID
- 1390852746524311296
-
- NII論文ID
- 230000012765
-
- ISSN
- 21885079
-
- 本文言語コード
- en
-
- データソース種別
-
- JaLC
- CiNii Articles
-
- 抄録ライセンスフラグ
- 使用不可