@techreport{oai:ipsj.ixsq.nii.ac.jp:00232583,
 author = {上野, 貴弘 and 大園, 倖暉 and 大橋, 正良 and Takahiro, Ueno and Koki, Ozono and Masayoshi, Ohashi},
 issue = {23},
 month = {Feb},
 note = {現在我々は，福岡県直方市内の遠賀川流域に設置された樋門の遠隔監視・制御に取り組んでいる．樋門制御には，重要な判断要素として支川の逆流がある．そのため，支川の水向を予測する機械学習モデルの開発が急務となる．そこで，圧力センサと監視カメラによる各種データを組み合わせた水向予測に焦点を当てる．圧力センサは水中の水向データを取得し，監視カメラは川の流向を表現した時系列画像データを取得している．水向予測に向けて，支川で計測を行い，取得データに基づき機械学習モデルを開発した．本稿では，モデルの評価結果に加えて，データ計測環境の調査結果について報告する．, Our laboratory is engaged in the research and development of a system for the remote control and monitoring of sluice gates in the Onga River that flows through Nogata City. The backflow from tributaries is a significant factor for sluice gate control decisions. Development of a machine learning model to predict the direction of water flow in these tributaries is essential. We focus on predicting water flow direction by combining data obtained from pressure sensor and surveillance camera. Pressure sensor collects data on underwater flow direction, while surveillance camera provides time-series image data of flow direction within the water. To predict water flow, we conducted measurements in the tributaries and developed the machine learning model using the collected data. In this paper, we report the evaluation results of the model and insights from an investigation into the data measurement environment.},
 title = {遠賀川流域における樋門遠隔制御のための水向予測モデルの開発と評価},
 year = {2024}
}