Abstract
Today, several soccer teams make use of the capabilities provided by analysis systems to further improve the overall team and player performance.
As more and more analysis systems are developed, using these types of tools have become common and the advantages these systems provides can be the difference between winning or loosing a match.
To offer an analysis system that can capture team data, annotate important events and deliver real-time video, we present the idea of Bagadus. With all of the these goals fulfilled, it is able to provide all the capabilities needed for team development. The system combines player tracking sensors, a easy-to-use annotation system and cameras to capture video of the games. The goal of Bagadus is to be fully automatic and real-time, but currently there are several issues that needs to be solved before one can take full advantage of the system.
In this thesis, we present an improved version of the previously existing Bagdus video capture component. By designing a pipeline for panorama creation, we are able to deliver a panorama video in real-time consisting of frames from four individual cameras.
We will also look further into the visual output of the pipeline and how we can improve the image quality by applying a method for color correction.