The Bagadus system is a largely automated tool designed to improve athlete performance by combining subsystems for sports analytics, player tracking and video capture, and is currently installed in several Norwegian soccer stadiums where it is used daily by coaches to observe and evaluate the performance of their athletes. It does this efficiently by automating the process of integrating these subsystems and their data. One important part of this system is the video capture subsystem which creates real-time panoramic videos from an array of HD cameras overlooking the field. The output of this image stitching pipeline is read by a panoramic video viewer application which takes the panoramic video and corrects for the image warping which results from the panorama projection process. This allows users to view the panoramic video through a virtual camera with features such as straight lines preserved. This camera is freely controllable by the user, who may zoom and pan the video in real-time. The existing implementation of this viewer relies on the CUDA parallel computing platform, a closed proprietary technology requiring a discrete GPU from the same vendor. In this thesis we evaluate this viewer, and present a prototype of an alternative panoramic video viewer that uses open web standards such as WebGL. This prototype is capable of generating a virtual camera that interacts with high resolution panoramic video with equivalent graphical fidelity to that of the previous implementation on a much broader set of hardware and with far greater ease of use and deployment. We also introduce multiple improvements to the user interface and control methods, such as gamepad support, as well as techniques for increasing performance and reducing bandwidth consumption by segmenting and adapting the video stream based on how the virtual camera is observing the panoramic video.