Date of Award

12-2025

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Computer Science

Committee Chair/Advisor

Dr. Jerry Tessendorf

Committee Member

Dr. Federico Iuricich

Committee Member

Dr. Daljit Singh Dhillon

Committee Member

Dr. Matias Volonte

Abstract

An intersection detection method between analytical descriptions of curved rays and geometric primitives is implemented to facilitate the process of rendering scenes containing a medium of continuously variable index of refraction (IOR) via a ray marching framework. Two KD Tree designs are introduced that utilize the curved ray intersection detection technique to accelerate the process of rendering scenes containing many triangles. A method for sub-sampling steps in the ray march is introduced to facilitate rendering scattering volumes with curved rays.

Three scenes containing physically-based IOR gradient profiles conducive to the formation of optical phenomena are constructed to demonstrate the rendering capabilities of this intersection detection method inside of a curved ray marching framework. These scenes consist of either an inferior mirage, superior mirage, or atmospheric turbulence. Mirages are generated by modulating the temperature profile of the atmosphere, and turbulence is generated using a C2 n model of turbulence. Atmospheres with Rayleigh scattering are implemented in mirage scenes, and sunsets are modeled through the usage of Deep Shadow Maps. Renders are evaluated qualitatively by reviewing the visual artifacts present in the render when adjusting the tolerance parameters for detecting an intersection and for presuming that the IOR remains constant for a single step of the march.

Renders are evaluated quantitatively by calculating the amount of speedup gained by using the modified KD Trees to sort scene geometry, and by altering the aforementioned tolerance parameters.

Author ORCID Identifier

0009-0003-8591-0598

Available for download on Thursday, December 31, 2026

Share

COinS