Optical tracking in team sports

Abstract Sports analysis has gained paramount importance for coaches, scouts, and fans. Recently, computer vision researchers have taken on the challenge of collecting necessary data by proposing several methods automatic player ball tracking. Building gathered tracking data, miners are able to perform quantitative performance players teams. With this survey, our goal is provide a basic understanding analysts about process creating input characteristics thereof. Thus, we summarize recent optical providing comprehensive taxonomy conventional deep learning methods, separately. Moreover, discuss preprocessing steps tracking, most common challenges in domain, application sports Finally, compare their cost limitations, conclude work highlighting potential future research directions.