As an important component of China’s intangible cultural heritage, the standardization of movements in traditional martial arts is the key to the inheritance of the art. In this study, a motion capture analysis system for traditional martial arts is designed with Kinect motion capture model as the core, combined with Unity3D technology and Dynamic Time Warping (DTW) algorithm. The system can capture human motion data stored in a standard database by Kinect device and provides videos related to traditional martial arts learning. Combined with the DTW algorithm to match the joint angle and velocity characteristics during movement, the automatic scoring of traditional wushu movements was realized. The average three-dimensional motion range of the spine in traditional wushu is between 137±13°, and the average range on two-dimensional recognition is between 138±17°, with better recognition effect on different levels. The correlation between the 3D coordinate-time curves of human joint points obtained by the system analysis of this paper and the manual analysis is strong, and the correlation coefficients are all greater than 0.90. The traditional wushu movement curves measured by this system are basically consistent with the actual movement curves. In addition, on the scoring of traditional wushu movements, the error between the scoring of this system and the scoring of experts was between ±0.2 points. The results of this study promote the digital inheritance of the intangible cultural heritage of sports and have practical value for the dissemination of traditional martial arts.
