Matthew Jung Undergraduate Dissertation 2014/15

Real-Time Image Stitching of UAV Cameras for the Oculus Rift

Supervised by P.Richmond

Abstract

Drones or unmanned aerial vehicles (UAVs) are fast becoming a centerpiece of new technology, much of which has to do with their untapped potential. Their uses are vast and include aerial surveillance, scientific research, search and resue, military operations and even filmmaking. Another hugely popular phenomenon is the renewed interest and development of virtual reality head sets. Combining these two technologies could result in a system which could virtually place a user on board a UAV. The experience would be tactile and functional, allowing users to respond to situations intuitively with head gestures. With the help of the Oculus Rift, this project aims to create a proof of concept of this very idea. To do this we will employ the widely used images stitching technique to create  panorama from smeller constituent images to reproduce a 360 degree panorama. The most difficult part of this project will be to deliver a system which is able to stitch multiple camera feeds and to create a virtual environement viewable on the Oculus Rift, all in real-time. Therefore one of the most important aspects which will be looked at is dealing with issues which could slow down the frame rate and the unique strategies to combat these effects.