Abstract:
The goal for the image processing software is to perform Detection of each basketball player on the court, Team classification, and Individual player tracking. We first plan to detect each individual player on the court while minimizing the detection of non-players (crowd and referees). The video frames of the game will come from a static camera view that will not change. Next, we’ll classify each player as either Team A or Team B based on a unique identifier, such as jersey color. Finally, we will try to maintain a track on each player so that their position can be stored and analyzed. Difficulties may arise as players move around each other on the court and cross paths. A stretch goal includes displaying colored dots overlaid onto a court image that shows each player’s team and current position, as well a statistics based on each player’s location. The player detection and tracking software will be implemented in MATLAB 2013a (Mathworks). we were able to accurately detect and track the individual players until more complex situations arose, such as players overlapping on the court. In ideal situations, these techniques provided reliable detection and tracking.
Author Name : R.Shakila, Dr. G.Nallavan
Keywords: Tracking, Image Processing, Basketball, Classification, Homography Technique.
Leave a Reply