Date of Award


Document Type


Degree Name

Master of Science


Department of Engineering Physics

First Advisor

Samuel Butler, PhD


Accurate Bidirectional Reflectance Distribution Function (BRDF) models provide critical scatter behavior for computer graphics and remote sensing performance. The popular microfacet class of BRDF models is geometric-based and computationally inexpensive compared to wave-optics models. Microfacet models commonly account for surface scatter and Lambertian volume scatter, but not directional volume scatter. This work proposes directional volume scatter modeling for enhanced performance over all observation regions. Five directional volume models are incorporated into the modified Cook-Torrance microfacet model. Additionally, a semi-empirical directional volume term is presented based on the Beckmann microfacet distribution and a modified Fresnel reflection term. High fidelity, low density data from 15 datasets are fit to each hybrid model using a recursive optimization method then compared to the baseline Cook-Torrance model. By including a directional volume term, analysis shows fit quality is improved based on the square of the mean standard error (MSE2) by as much as 78% and backscatter agreement is improved by as much as 92%. Including the semi-empirical, Oren-Nayar, or Beard-Maxwell directional volume term reduced backscatter MSE2 across datasets exhibiting high volume scatter by an average of 52%, 46%, and 26% respectively. Directional volume terms showed statistically insignificant improvement for low volume scatter materials, while full model improvements were apparent across all high volume scatter visually diffuse materials. Results suggest directional volume scatter modeling can consistently improve full model fit quality with emphasized model agreement for backscatter observations. These results validate directional volume scatter significance and are expected to lead to enhanced remote sensing and scene generation.

AFIT Designator


DTIC Accession Number