Efficient implementation and processing of a real-time panorama video pipeline with emphasis on dynamic stitching

Abstract

The Bagadus system has been introduced as an automated tool for soccer analysis, and it is built up by an analysis subsystem, tracking subsystem and video subsystem. Bagadus allows for simplified soccer analysis, with the goal of improving athletes' performance, by automating the integration of these subsystems. The system is currently installed at Alfheim stadium in Tromsø, Norway. An important part of the video subsystem is the creation of panorama videos from four HD cameras. However, the stitching pipeline for panorama video generation in the first version of the system did not manage to do this in real-time.

In this thesis, we present how to build an improved panorama stitcher pipeline that is able to stitch video from four HD cameras into a panorama video in real-time. We will detail the architecture and modules of this pipeline, and analyze the performance. In addition we will focus on the stitching component, and how that can improve the overall visual quality of the output panorama.