Dárkon

Abstract

Sports today are more competitive than ever, with an eye for extreme details. Small margins can differentiate between succeeding or failing. That is why it is essential for a team to pay attention to all these margins and use them to their advantage. Technology today plays a huge part in sports. With video cameras and sensors following the players’ every move, it is vital to use this technology to gain an advantage over the opponent.

This thesis presents Dárkon, a video analysis system to help teams with their video analysis needs. Dárkon works together with a tagging system developed simultaneously to provide filtering based on field position, type of event, and outcome of event to find a video of a specific event or playlist of events. Along with this, Dárkon will attempt to collect videos from different sources, to bring everything into one coherent system.

The videos of a single event, sequence, or playlists are generated on the fly from the metadata for the entire video to keep the storage to a minimum, while still being able to only play specific parts of a video from an entire match.