Algorithmic Model of Microfacet BRDF for Monte Carlo Calculation of Optical Radiation Transfer
Alexander V. Prokhorov, Leonard M. Hanssen · 2003
An algorithmic model of bi-directional reflectance distribution function (BRDF) based on the ray optics approximation and microfacet model of randomly rough surface is proposed. Its central idea is that for every incident ray, the normal vector to the surface undergoes a random perturbation, and the direction of specular reflection is calculated using this perturbed normal. Such behavior of the normal can be treated within the framework of a microfacet model of randomly rough surfaces. The algorithm allows one to model reflection from both isotropic and anisotropic surfaces, with twodimensional Gaussian and other probability density functions for the normal vector perturbations, and various geometrical attenuation functions. The proposed “perturbed normal microfacet (PNMF) model” exhibits experimentally observed effects such as increased reflectance near grazing incidence and off-specular peaks, and allows fast importance sampling. A weighted sum of Lambertian and PNMF BRDFs can be fitted to experimental data by varying the appropriate parameters. Adherence to the reciprocity principle and energy conservation law is demonstrated via results of forward and backward ray tracing. The PNMF model can be used in Monte Carlo calculations of radiative heat exchange among rough surfaces, in realistic image synthesis, lighting engineering, for modeling of such radiometric devices as blackbody radiation sources, integrating spheres in the infrared spectral range, cavity detectors of radiation, diffusely reflected panels, etc.