ABSTRACT
This chapter provides an overview of computer vision methods and addresses applications in team game analysis. One focus is on the automatic tracking of players and game objects (balls, pucks, etc.). Advantages and disadvantages compared to other technologies are shown. In addition to basic methods for object tracking and more sophisticated approaches based on machine learning, specific solutions in the case of occlusions, such as probabilistic models, are presented. Another section of the chapter shows how to validate such video-based systems. In particular, comparison systems based on infrared cameras are used as a gold standard. The last section of the chapter shows concrete application scenarios of video-based systems in the analysis of the physical load in competition and training as well as for the analysis of tactical behavior. Concrete developments for analysis and use in football, hockey, basketball and other team sports are shown as examples.
