The Intelligent Computing and Information Technology in Sports Performance Evaluation

Abstract

This paper presents a new method to obtain the training trajectory by using the mean shift method. In this way, the incomplete motion trajectory caused by the rapid movement of the moving target due to the complex background is solved. The human body modeling is regarded as a skeletal model with 51 degrees of freedom and 16 joints, and the motion trajectory is digitally processed. At the same time, the dimension compression of the trajectory is also carried out to reduce the calculation amount. The gradient iteration method based on random distribution is selected to reduce the dependence on environmental parameters. The object color image is the main feature to realize the acquisition of motion trajectory. The experiment proves that the algorithm can reflect the movement state of each part of the athlete well. This method can accurately obtain the training trajectory without any associated parameters.