A Sports Digital Training System Based on Middle and Bottom Visual Information

Abstract

Physical exercise has become increasingly popular in China in recent years, and the number of individuals participating in physical activities is growing year after year. Therefore, they are facing the problem of a serious shortage of sports resources, especially in the central and western regions where a complete physical education teaching system has not been established yet. The introduction of sports digital training alleviates this problem to a great extent. This approach uses advanced computer technology for digital training, collects people's sports information, and formulates suitable sports programs for different groups. The system can collect people's action videos at the same time by two cameras, analyze the athletes' actions from the side and face by using the software, and compare the professional athletes' actions with the collected video action for in-depth analysis. The sports digital training system proposed in this study focuses on collecting the bottom visual information of sports. The visual information in the middle and low levels is mainly used to describe the content of video segments, i.e., key primitives, including motion trajectory, motion action, and tracking motion, and establishes a sports digital training system based on the bottom visual information. In addition to the completion of basic functions of a sports digital training system, the system also adds the functions of video segmentation, action freezing, speed calculation, table display, local locking, and tracking, which can accurately collect the bottom visual information of people's sports and realizes the digital sports training.