Multi-Device Environment Reconstruction for Augmented Reality Collaboration

Kurzfassung

In the modern technological age, the ability to connect and collaborate with people in all kinds of places are only increasing. Collaboration between experts of many fields is thus becoming more and more relevant. Modern Augmented Reality devices provide an immersive experience for this kind of collaborative work and the possibilities only increase as the time goes on. In many fields that require high precision like, for example, the field of engineering, a consensus about the topic that is talked about is important to an elementary degree. For that, the display of models of the object of consideration has been a basic part of these fields for a long time now. The field of photogrammetry is still in rapid development and has shown that objects can be reconstructed based on a set of images taken of it. This has made creating 3D scans of objects more available, since it does not need to rely on expensive scanners anymore. These developments enable the use of Augmented Reality devices to be used as displays and scanners for Collaborative viewing. In this thesis, the feasibility of using Augmented Reality devices for the purpose of scanning these scenes is tested, as well as their ability to scan with multiple devices at once, while enhancing the results of the scan. For that Structure from Motion and Multi View Stereo have been implemented as methods of photogrammetry. A prototype was developed to test the performance of Structure from Motion and Multi View Stereo for this purpose. The prototype utilizes the scanners of the HoloLens to aid the reconstruction and further improvements to this method are explored based on the data, collected in the tests.