A Geometric Algorithm for Snow Distribution in Virtual Scenes

Abstract

Freshly fallen snow is a popular natural phenomenon able to evoke great beauty in all kinds of scenes. However, there still does not exist an all-purpose algorithm for automated snow distribution in virtual worlds. Previous works modelled snow either relying on costly particle simulations or oversimplified surface displacements. In this paper we develop a novel geometric snow model. In a first step, we propose a statistical snow deposition model inspired by real world observations. This statistical model is used to derive a geometric snow shape formulation. The geometric scheme is implemented with an enhanced height span map. Scene geometry is expressed with two parameter fields that enable us to efficiently compute snow cover geometry. A selection of snow covered scenes demonstrates the realism that can be achieved with this new method.