Identifying and following players in Soccer Track footage

Abstract

Improving both individual and team performance is possible via analysis of player movement on the field during a match. Current methods frequently use cumbersome GPS trackers strapped to players or a network of several cameras, each of which monitors a different area of the field, to monitor how the football players move about. This research introduces a new method that utilizes cutting-edge object recognition and tracking algorithms in conjunction with a single high-resolution fisheye camera. At all times, the suggested technique gives you the right player identifiers and locations. We tested three object tracking algorithms and investigated many variations of the YOLOv8 algorithm that used different picture pre-processing methods. This method drastically cuts down on system costs, allowing even smaller teams to use precise match data to their advantage.