Today we live in a world where businesses still opt to send employees on flights to the other side of the world, rather than make use of existing telepresence applications. It is believed that the reluctance to use existing applications is largely due to the fact that they do not make the users feel truly present. With the current improvements being made in VR and AR technologies, the opportunity exists to create a new form of telepresence that allows a user to place themselves in a virtual meeting space with increased potential for communication. Existing solutions for geometrically reconstructing dynamic environments require calibrated camera arrays which are difficult to setup. This thesis proposes research into developing a SLAM based dynamic scene reconstruction system for a novel hybrid camera system, which reconstructs both rigid and non-rigid objects in real-time. Rigid and non-rigid SLAM based reconstruction approaches exist, but none of these methods provide a full solution. The preliminary work has provided a method of adapting existing approaches to our novel camera setup, as well as supplying an improved tracking algorithm which will be used over the remainder of the PhD to expand upon existing works and provide a full dynamic geometric reconstruction algorithm.