Development of an Exercise Posture Monitoring System for Air Cadets by Using MediaPipe and OpenCV

This paper addresses the challenge of standardizing the pulling posture test at the Air Force School, a crucial element in evaluating physical fitness and a key factor in selecting aviation cadets. The existing problem stems from the lack of standardization, as the scoring of the pulling posture test relies on the subjective judgment of the evaluation committee. To tackle this issue, this paper proposes a systematic approach to standardize and establish consistent criteria for the pulling posture test. We introduce a novel system for assessing students' pulling postures, leveraging MediaPipe and OpenCV technologies. Our system involves programming and utilizing MediaPipe and OpenCV modules to access the camera and detect facial features, hand movements, and angles of both arms. When a participant performs the pulling motion, the system evaluates whether it meets predefined criteria. Successful attempts trigger an auditory confirmation, and the system automatically records and logs the count. To validate the accuracy of our system, we conducted tests in which instructors manually counted alongside the automated system. The results indicate the system's capabilities to accurately detect pulling postures, although further development is necessary to address challenges related to camera height and angle. Our findings suggest that this system holds promise in standardizing the pulling posture test, providing a valuable tool for the selection of aviation cadets at the Air Force School.