In this thesis, the theory of video seethrough is fundamentally presented on the basis of a panoramic view from several camera frames of different perspectives. Based on this, a system was designed and implemented in which video streams are put together into a panoramic image by perspective distortion. This is then projected onto the inside of a cylinder with the virtual position of the viewer in the middle. Finally, the resulting video panoramas will be displayed in VR glasses. Within the implementation some optimizations are also presented, among others those that make the system real-time capable beyond the task. Furthermore, the developed system will be evaluated and compared with two other methods.