Multi-Scale Radiative Transfer Simulation for the Scattering of Light by Microgeometry

Light is electromagnetic energy propagating through space as waves, and if the light does not intersect with matter, the light will continue to propagate
indefinitely. When the light does intersect matter, complex processes occur, which gives rise to considerable computational challenges if these processes are to be simulated. Additionally, the short wavelength of light makes the scattering of light affected by even small structures on a surface. The small structures can cause the light to change the direction of propagation many times or even cause the light to change its color. With the increasing complexity of surfaces and scattering, the required numerical computation increases substantially. It can become infeasible to obtain simulated results if all the details of light’s behavior are in the simulation. In this thesis, we make a range of assumptions to obtain faster simulation. We assume that light propagates in a relatively straightforward manner, and we limit the types of surfaces we support with our methods while at the same time providing a reasonable approximate result. We reduce the resolution of our computations and results if details are not visible for the human eye or, e.g., a camera.
We focus on surfaces that contain both microscopic and macroscopic features and how we may connect these features into one result. We exploit knowledge of scattering processes on a microscopic scale to infer information about the underlying microgeometry based only on a macroscopic result. We build models for and compute scattering of light on engineered surfaces. Specifically, we target surfaces with functionality where the scattering of light is controlled with the microstructure and surfaces, which cannot be represented using only a heightmap.