Application of computer vision to automate notation for tactical analysis of badminton

In tactical analysis of fast-paced net-based sports such as badminton, the identification of habits and movement of an opponent provides an advantage in terms of reaction time, which results in a player obtaining control of the game. A tactical analyst in badminton segments the court using imaginary lines, and intuitively notates the position of a player according to those segments throughout the game. This paper presents a computer vision based approach to automate this notational process. Motion tracking of the badminton players identifies their position throughout the game. Different court segmentation patterns are tested to identify the borders of the segments to match the intuitive segmentation performed by a badminton tactical analyst. By application of a dynamic window for assigning players to court segments, we achieve an accuracy level of above 85%.