Global Illumination Computation for Augmented Reality

In order to merge virtual objects into a real scene seamlessly, it is important to maintain consistent common illumination. In this thesis, we propose a framework that captures high dynamic range light probes and decomposes them into sets of directional light sources in real-time. The light sources, captured and processed on a dedicated device, are than made available to rendering engines via a server that provides wireless access. We implement three different importance sampling techniques and compare them in terms of the quality of sampling pattern, temporal coherence, and performance.