The construction of sports training quality evaluation model based on sensor data

Based on the visual quality of the training evaluation as a technology is closely related to People’s Daily life, has the very high research value, and in the near future will be in medical, health, security, etc. has a broad application prospect. In the human body based on visual training quality evaluation in the research of this field, human action recognition technology is one of the core technology, is also a research hotspot recently. The technology is mainly to solve the classification problem of human actions, but in practice, it is not enough to study classification problem, sometimes need to evaluate the quality of the completion of human action, namely human motion evaluation technology, provides the user with feedback, correct mistakes, thus improve the action degree of standardization. This paper analyzes and compares all kinds of action selection radio gymnastic sports training as a quality evaluation object, and using Microsoft’s depth camera device as acquisition equipment, data to obtain three-dimensional skeleton of the body joints. Human bone joint data from the device has the problem of “distortion”, this paper adopted the average filtering algorithm for data preprocessing, the experimental results show that average filtering algorithm can not only protect. The high sensitivity and good smoothing effect.