Research on sustainable fishery condition monitoring through cooperation between fishermen and artificial intelligence technology

2023 Fiscal Year Final Research Report

• Project/Area Number: 22K19218
• Research Category: Grant-in-Aid for Challenging Research (Exploratory)
• Allocation Type: Multi-year Fund
• Review Section: Medium-sized Section 40:Forestry and forest products science, applied aquatic science, and related fields
• Research Institution: Waseda University
• Principal Investigator: Tetsuji Ogawa 早稲田大学, 理工学術院, 教授 (70386598)
• Project Period (FY): 2022-06-30 – 2024-03-31
• Keywords: 半教師あり学習 / 深層ニューラルネットワーク / 海況シミュレーション / 漁場状態監視 / マルソウダ曳縄漁

Outline of Final Research Achievements

This study explored fundamental technologies and the design of monitoring interfaces to enable the assessment and explanation of fishing ground conditions using meteorological and oceanographic data. Specifically, we developed a technique that accurately narrows the prediction range of optimal fishing grounds, while preventing detection omissions of actual fishing grounds. This method does not restrict predictions to previously operated sea areas, and instead utilizes unlabeled data from unexplored regions through semi-supervised learning. The effectiveness of this approach was demonstrated in the bullet tuna trolling fishery in Tosashimizu City, Kochi Prefecture. Additionally, we created a web interface to display daily updated predictions of optimal fishing grounds, which were then shared with fishermen and fisheries experts.

Free Research Field

知覚情報処理，ヒューマンインタフェース，水圏生産科学

Academic Significance and Societal Importance of the Research Achievements

広大な未操業海域に対するデータの効率的利用は，漁獲に関するデータ収集が容易でない状況において良漁場を高精度に予測するための効果的な手段である．このように，ラベルなしデータを有効活用してシステムを効率的に構築・成長させる方法論の確立は，漁場の監視のみならず機械学習技術を社会実装するための本質的な課題であり，本研究を通じて得た予測技術およびその評価の枠組みに関する知見は広く学術的意義がある．また，漁場の高精度な監視・予測技術は，気象・海況情報を手掛かりに経験と勘に頼って意思決定を行っている現在の漁業の在り方を，データに基づく資源管理型の漁業に転換する足掛かりとなり得るもので，社会的意義も大きい．