Research attempting non-contact, real-time estimation of metabolic rate using AI from thermal images for air conditioning control

2023 Fiscal Year Final Research Report

• Project/Area Number: 20K04824
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 23020:Architectural environment and building equipment-related
• Research Institution: National Institute of Technology (KOSEN), Kure College
• Principal Investigator: Yamato Yoshiaki 呉工業高等専門学校, 建築学分野, 教授 (20450140)
• Co-Investigator(Kenkyū-buntansha): 宋 城基 広島工業大学, 環境学部, 教授 (70422813) ; 木村 竜士 高知工業高等専門学校, ソーシャルデザイン工学科, 准教授 (90571810)
• Project Period (FY): 2020-04-01 – 2024-03-31
• Keywords: 代謝量 / 姿勢 / 物体検知AI / 姿勢推定AI / YOLO / POSENET / オープンデータセット

Outline of Final Research Achievements

The purpose of this research is to develop a system that uses AI to determine the posture and movement of the human body in thermal images, and to estimate metabolic rate in real time and non-contact. Attempt was made to determine the pose of person in thermal images, using object detection AI and pose estimation AI. In order to facilitate the collection of various images of person in different poses, which are essential for AI-based pose estimation, the effectiveness of online image data and additional open datasets for training was also considered, and useful datas were gained. Furthermore, the metabolic equivalent (METs) for various postures and movements based on experiments with human subjects were presented . Relationship between heart rate and the metabolic equivalent were also presented.

Free Research Field

建築環境工学

Academic Significance and Societal Importance of the Research Achievements

体感温度に影響する気温・湿度・気流・放射・代謝量・着衣量の温熱6要因のうち，これまでになおざりにされていた人体側2要因のひとつである代謝量のリアルタイム・非接触な推定方法を，特に様々な分野で導入され，導入された分野でより一層の技術革新を呼ぶ相乗効果を生み出しているAIを用いて試み，その可能性と課題を示した．
得られた知見は，今後に計画している着衣量のリアルタイム・非接触推定にも生かされ，これらは健康的で快適でありながらより省エネな空調制御技術の発展に貢献する．