• Search Research Projects
  • Search Researchers
  • How to Use
  1. Back to previous page

機械学習を用いた極域における熱輸送メカニズムの解明

Research Project

Project/Area Number 20K11718
Research Category

Grant-in-Aid for Scientific Research (C)

Allocation TypeMulti-year Fund
Section一般
Review Section Basic Section 60030:Statistical science-related
Research InstitutionHosei University

Principal Investigator

伊藤 香寿恵 (鈴木香寿恵)  法政大学, 理工学部, 助手 (20455190)

Co-Investigator(Kenkyū-buntansha) 山内 恭  国立極地研究所, その他部局等, 名誉教授 (00141995)
徳永 旭将  九州工業大学, 大学院情報工学研究院, 准教授 (50614806)
Project Period (FY) 2020-04-01 – 2025-03-31
Project Status Granted (Fiscal Year 2023)
Budget Amount *help
¥4,420,000 (Direct Cost: ¥3,400,000、Indirect Cost: ¥1,020,000)
Fiscal Year 2022: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000)
Fiscal Year 2021: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000)
Fiscal Year 2020: ¥1,820,000 (Direct Cost: ¥1,400,000、Indirect Cost: ¥420,000)
Keywords半自動学習 / 衛星観測 / Atmopheric River / Atmospheric River / 機械学習 / 時系列解析 / 熱収支 / 熱輸送 / 地球温暖化
Outline of Research at the Start

衛星観測による雲画像は,上空から輝度温度を測定した結果であり,雲の三次元構造にそのまま置き換えることは難しい。
雲構造の解釈のために気象数値モデルを使って時間変化も含めた雲の四次元構造を捉え,雲発達を予測する統計モデルを構築する。 雲発達モデルを未学習の雲画像に対して適用させ,降雪をもたらす雲が発生した時点で自動検出することが目標である。 また,両極域を対象とした解析や予測を実施し,南北における熱輸送の違いから,地球温暖化の影響について検討する。

Outline of Annual Research Achievements

これまで作成した衛星観測雲画像のうち,Atmospheric Riverとした雲について作成した正例・負例データセットを用いた教師データを減らして学習する手法について検証を進めた.CoSPAは物体識別の際に必要となる画像セグメンテーションを半教師あり学習という形で行う新しい画像処理のための機械学習の手法であり,本研究で用いている衛星観測雲画像に対して適用を試みた.CoSPAの有用性を検証するために,一般的なオートエンコーダとしてU-Netによるセグメンテーションも実行した.
セグメンテーションの評価指標として,ここではIoU(Intersection over Union)とDiceの平均値をデータセット毎に算出した.結果から,CoSPAによるセグメンテーション生成は一般的なU-Netより優れていると判断できるが,ピクセル単位での評価であることに注意が必要である.したがって,手動で教師データのタグ付けを行うよりもある程度スクリーニングを実施することが出来ると予想できる.今後は,新たに入手する衛星観測雲画像に対して,CoSPAによる教師データ生成を実施し,最終的に人間による手動判別を実施することで教師データ生成が容易になると期待している.
また,観測データの理解を深めるため,雪のサンプリング観測に参加した.研究成果の解釈や,実データへの適用に役立つ情報を得ることができ,最終年度につなげていく予定である.

Current Status of Research Progress
Current Status of Research Progress

3: Progress in research has been slightly delayed.

Reason

北極側のデータセット作りが遅れているが,正例を自動的に生成するための深層学習は概ね準備が整ってきた.

Strategy for Future Research Activity

研究代表者の所属先が変わり,研究に割ける時間が大幅に増加した.再延長し,最後のデータセット作りを迅速に行い,両極の雲画像について自動識別を実施する.結果をまとめて学会発表および雑誌論文として報告する.

Report

(4 results)
  • 2023 Research-status Report
  • 2022 Research-status Report
  • 2021 Research-status Report
  • 2020 Research-status Report
  • Research Products

    (22 results)

All 2024 2023 2022 2021 2020 Other

All Int'l Joint Research (2 results) Journal Article (5 results) (of which Open Access: 4 results,  Peer Reviewed: 3 results) Presentation (14 results) (of which Int'l Joint Research: 12 results) Book (1 results)

  • [Int'l Joint Research] ウィスコンシン大(米国)

    • Related Report
      2021 Research-status Report
  • [Int'l Joint Research] ウィスコンシン大学(米国)

    • Related Report
      2020 Research-status Report
  • [Journal Article] High-top clouds play an efficient part in moisture transport to the Antarctic2022

    • Author(s)
      Suzuki Kazue、Tokunaga Terumasa、Yamanouchi Takashi、Motoyama Hideaki
    • Journal Title

      Authorea

      Volume: -

    • DOI

      10.1002/essoar.10512276.1

    • Related Report
      2022 Research-status Report
    • Open Access
  • [Journal Article] Identifying Snowfall Clouds at Syowa Station, Antarctica via a Convolutional Neural Network2021

    • Author(s)
      Suzuki Kazue、Shimomura Masaki、Nakamura Kazuyuki、Hirasawa Naohiko、Yabuki Hironori、Yamanouchi Takashi、Tokunaga Terumasa
    • Journal Title

      Advances in Artificial Intelligence. JSAI 2020. Advances in Intelligent Systems and Computing

      Volume: 1357 Pages: 73-83

    • DOI

      10.1007/978-3-030-73113-7_7

    • NAID

      130007857079

    • ISBN
      9783030731120, 9783030731137
    • Related Report
      2021 Research-status Report
    • Peer Reviewed / Open Access
  • [Journal Article] Characterization of aerosol number size distributions and their effect on cloud properties at Syowa Station, Antarctica2021

    • Author(s)
      Hara Keiichiro、Nishita-Hara Chiharu、Osada Kazuo、Yabuki Masanori、Yamanouchi Takashi
    • Journal Title

      Atmospheric Chemistry and Physics

      Volume: 21 Issue: 15 Pages: 12155-12172

    • DOI

      10.5194/acp-21-12155-2021

    • Related Report
      2021 Research-status Report
    • Peer Reviewed / Open Access
  • [Journal Article] Arctic warming amplification and warming suppression in East Antarctica - Contribution of MOC to north-south asymmetry -2021

    • Author(s)
      Takashi Yamanouchi
    • Journal Title

      Okhotsk Sea and Polar Oceans Research

      Volume: 5 Pages: 1-6

    • NAID

      40022791134

    • Related Report
      2020 Research-status Report
    • Peer Reviewed
  • [Journal Article] Identifying the Snowfall Cloud at Syowa Station, Antarctica via a Convolutional Neural Network2020

    • Author(s)
      Kazue SUZUKI, Masaki SHIMOMURA, Kazuyuki NAKAMURA, Naohiko HIRASAWA, Hironori YABUKI, Takashi YAMANOUCHI, Terumasa TOKUNAGA
    • Journal Title

      Proceedings of the Annual Conference of JSAI

      Volume: JSAI2020 Issue: 0 Pages: 3F1ES205-3F1ES205

    • DOI

      10.11517/pjsai.JSAI2020.0_3F1ES205

    • NAID

      130007857079

    • Related Report
      2020 Research-status Report
    • Open Access
  • [Presentation] Heavy Snow Cloud Detection in Satellite Images Based on Semi-Supervised Image Segmentation2024

    • Author(s)
      Lin Magari, Terumasa Tokunaga, Kazue Suzuki
    • Organizer
      The 5th International Symposium on Neuromorphic AI Hardware
    • Related Report
      2023 Research-status Report
    • Int'l Joint Research
  • [Presentation] Heavy Snow Cloud Detection in Satellite Images Based on Semi-Supervised Image Segmentation2023

    • Author(s)
      Lin Magari, Terumasa Tokunaga, Kazue Suzuki
    • Organizer
      日本地球惑星科学連合2023年大会
    • Related Report
      2023 Research-status Report
    • Int'l Joint Research
  • [Presentation] An attempt to detect of Atmospheric River by Satellite Cloud Images in the Antarctic2023

    • Author(s)
      Kazue Suzuki, Terumasa Tokunaga, Naohiko Hirasawa, Takashi Yamanouchi, Hideaki Motoyama
    • Organizer
      The 14th Symposium on Polar Science
    • Related Report
      2023 Research-status Report
    • Int'l Joint Research
  • [Presentation] Heavy Snow Cloud Detection in Satellite Images Based on Semi-Supervised Image Segmentation2023

    • Author(s)
      Lin Magari, Terumasa Tokunaga, Kazue Suzuki
    • Organizer
      The 14th Symposium on Polar Science
    • Related Report
      2023 Research-status Report
    • Int'l Joint Research
  • [Presentation] A sequential trajectory method with data assimilation2023

    • Author(s)
      Kazue Suzuki, Yoshihiro Tomikawa
    • Organizer
      日本地球惑星科学連合2023年大会
    • Related Report
      2023 Research-status Report
    • Int'l Joint Research
  • [Presentation] Snowfall Events and Aerosol Atmospheric Rivers with Changes in Concentrations of Trace Substances in the Atmosphere in the Polar Regions2022

    • Author(s)
      鈴木 香寿恵、中村 和幸、徳永 旭将、原 圭一郎、後藤 大輔、平沢 尚彦、山内 恭
    • Organizer
      日本地球惑星科学連合2022年大会
    • Related Report
      2022 Research-status Report
    • Int'l Joint Research
  • [Presentation] Detection of Atmospheric River by Satellite Cloud Images in the Antarctic2022

    • Author(s)
      Kazue Suzuki, Terumasa Tokunaga, Naohiko Hirasawa, Takashi Yamanouchi, Hideaki Motoyama
    • Organizer
      The 17th Workshop on Antarctic Meteorology and Climate
    • Related Report
      2022 Research-status Report
    • Int'l Joint Research
  • [Presentation] The high-top cloud plays an efficient part in the moisture transport to the Antarctic2022

    • Author(s)
      Kazue Suzuki, Terumasa Tokunaga, Takashi Yamanouchi, Hideaki Motoyama
    • Organizer
      4th International Atmospheric Rivers Conference
    • Related Report
      2022 Research-status Report
    • Int'l Joint Research
  • [Presentation] 南極域における衛星雲画像による Atmospheric Riverの検出2022

    • Author(s)
      鈴木香寿恵, 徳永旭将, 平沢尚彦, 山内 恭
    • Organizer
      日本気象学会2022年度秋季大会
    • Related Report
      2022 Research-status Report
  • [Presentation] Identifying Snowfall Clouds at Syowa Station, Antarctica via a Convolutional Neural Network2021

    • Author(s)
      Suzuki Kazue、Shimomura Masaki、Nakamura Kazuyuki、Hirasawa Naohiko、Yabuki Hironori、Yamanouchi Takashi、Tokunaga Terumasa
    • Organizer
      16th Workshop on Antarctic Meteorology and Climate (WAMC, Virtual)
    • Related Report
      2021 Research-status Report
    • Int'l Joint Research
  • [Presentation] Atmospheric River による南極域へのエアロゾル輸送(1)2021

    • Author(s)
      鈴木香寿恵, 原圭一郎, 徳永旭将, 後藤大輔, 平沢尚彦, 山内 恭
    • Organizer
      日本気象学会2021年度秋季大会
    • Related Report
      2021 Research-status Report
  • [Presentation] Identifying the Snowfall Cloud at Syowa Station, Antarctica via a Convolutional Neural Network2020

    • Author(s)
      Kazue SUZUKI, Masaki SHIMOMURA, Kazuyuki NAKAMURA, Naohiko HIRASAWA, Hironori YABUKI, Takashi YAMANOUCHI, Terumasa TOKUNAGA
    • Organizer
      The 34th Annual Conference of the Japanese Society for Artificial Intelligence, 2020
    • Related Report
      2020 Research-status Report
    • Int'l Joint Research
  • [Presentation] The Identification of the snowfall cloud at Syowa Station, Antarctica2020

    • Author(s)
      Kazue SUZUKI, Masaki SHIMOMURA, Kazuyuki NAKAMURA, Naohiko HIRASAWA, Hironori YABUKI, Takashi YAMANOUCHI, Terumasa TOKUNAGA
    • Organizer
      AGU Fall Meeting 2020
    • Related Report
      2020 Research-status Report
    • Int'l Joint Research
  • [Presentation] Warm-moist air intrusion into the polar regions enhancing cloud longwave radiation and contributing to the warming2020

    • Author(s)
      Takashi Yamanouchi and Naohiko Hirasawa
    • Organizer
      Scientific Committee on Antarctic Research SCAR-OSC2020
    • Related Report
      2020 Research-status Report
    • Int'l Joint Research
  • [Book] Advances in Artificial Intelligence, Selected Papers from the Annual Conference of Japanese Society of Artificial Intelligence (JSAI 2020), L11: Identifying the Snowfall Cloud at Syowa Station, Antarctica via a Convolutional Neural Network2021

    • Author(s)
      Yada, K., Katagami, D., Takama, Y., Ito, T., Abe, A., Sato-Shimokawara, E., Mori, J., Matsumura, N., Kashima, H. (Editors), Kazue SUZUKI, Masaki SHIMOMURA, Kazuyuki NAKAMURA, Naohiko HIRASAWA, Hironori YABUKI, Takashi YAMANOUCHI, Terumasa TOKUNAGA (Coauthors)
    • Total Pages
      390
    • Publisher
      Springer International Publishing
    • ISBN
      9783030731137
    • Related Report
      2020 Research-status Report

URL: 

Published: 2020-04-28   Modified: 2024-12-25  

Information User Guide FAQ News Terms of Use Attribution of KAKENHI

Powered by NII kakenhi