Real-time modeling and rendering of raining scenes

Abstract

Real-time modeling and rendering of a realistic raining scene is a challenging task. This is because the visual effects of raining involve complex physical mechanisms, reflecting the physical, optical and statistical characteristics of raindrops, etc. In this paper, we propose a set of new methods to model the raining scene according to these physical mechanisms. Firstly, by adhering to the physical characteristic of raindrops, we model the shapes, movements and intensity of raindrops in different situations. Then, based on the principle of human vision persistence, we develop a new model to calculate the shapes and appearances of rain streaks. To render the foggy effect in a raining scene, we present a statistically based multi-particles scattering model exploiting the particle distribution coherence along each viewing ray. By decomposing the conventional equations of single scattering of non-isotropic light into two parts with the physical parameter independent part precalculated, we are able to render the respective scattering effect in real time. We also realize diffraction of lamps, wet ground, the ripples on puddles in the raining scene, as well as the beautiful rainbow. By incorporating GPU acceleration, our approach permits real-time walkthrough of various raining scenes with average 20 fps rendering speed and the results are quite satisfactory.