Development of a hybrid algorithm based on neural networks and classic methods of detection and tracking of fast-moving objects on an example a badminton shuttlecock

Michał Kopania

supervisor: Artur Przelaskowski

For the past few years, together with another Ph.D. student Jarosław Mówisz we have been building an instant review system for badminton. The system uses 14 to 20 fast cameras to track the shuttlecock and detect when and where it landed on the floor of a court. In order to track the object, it must be first found on an image among other objects. In a poster, I present a very fast, novel method I have developed for finding a fast-moving shuttlecock within the image from a camera.