Accurate Recognition of Volleyball Motion Based on Fusion of MEMS Inertial Measurement Unit and Video Analytic
Kaiqiao Peng, Yuliang Zhao, Xiaopeng Sha, Wenqian Ma, Yufan Wang, Wen J. Li · 2018
This paper presents an motion recognition and analysis method based on the fusion of MEMS Inertial Measurement Unit (IMU)data and the recorded video for the volleyball skill assessment. Based on the synchronous video and acceleration data of the player's wrist, the complete spiking motion are carefully studied. Six kinds of the important characteristics were used to recognize the 12 trails of spiking motion of the same player by neural network. The results demonstrated that this method can recognize the motions of the same players with an accuracy of 89.6%. Furthermore, this proposed method can be extended to analyze the key motions of other types of sports, e.g., basketball, badminton, and baseball.