Research on tennis video target tracking based on visual analysis

Abstract

With the advent of the era of artificial intelligence, great changes have taken place in the traditional broadcasting of sports games. As early as the early 20th century, the Hawkeye system appeared in tennis as a referee aid. However, there is currently no technology that can perform semantic analysis of live video of tennis matches. In the field of machine vision, target tracking is one of the most basic and important branches. It includes image processing, machine learning, pattern recognition and other aspects. It is a research direction involving a very wide range of fields. The complexity of the environment where the moving objects are located and the uncertainty of the moving objects themselves bring great challenges to the detection and tracking of moving objects. Traditional detection and tracking methods based on manual feature extraction have poor generalization ability and can not meet the needs of target detection and tracking in complex moving scenes. By estimating the two-dimensional posture of the players in the player area, analyzing their body skeleton and key information, the analysis of the players' movement types is completed. At the same time, according to the displacement information of key points of athletes' bodies and the running time of the system, the movement distance and speed of athletes can be obtained.