Real-Time Shadows for Large-Scale Geospatial Visualization

Michaela Niedermayer

Supervisor(s): J├╝rgen Waser, Daniel Cornel

VRVis Research Center

Abstract: Floods are often catastrophic and can cause enormous damage. Simulations are used to predict risks and respond to them early, like in the decision-support system Visdom. A true-to-life representation of the entire scene is important, as the visualization of the results must be understandable also to non-experts, such as decision-makers or the general public. What is missing so far are shadows, although they are particularly well suited to recognize relations of objects to each other. This paper covers the implementation of shadows in Visdom, to increase the realism of the scene. This is very complex for scenes on city or even country scale and requires a lot of own strategies, as there are no ready-made solutions. We present an adaptation of cascaded shadow maps for our purposes, as well as a variety of improvements to increase the shadow quality in our application. The result of this work is a flexible visualization of soft shadows for a variety of different-sized scenes in real time, which increase the realism and spatial perception.
Keywords: Real-time Graphics, Rendering
Full text:
Year: 2019