High speed ray tracing for a headlamp lens and advanced algorithms for ray analysis are investigated.
First, the basics of ray tracing, Algorithm to search intersection points between a ray and surfaces and refraction are reviewed, including intersection search for a ray with aspheric surfaces. A spherical surface, a plane surface, and a point cloud are reviewed as objects. Snell’s law is introduced from Fermat’s principle in 2D. Then, it extended to three dimensional spaces.
Second, photometry is reviewed for the post processing of ray tracing, due to the convolution effect of its area.
To accelerate ray tracing, the Nvidia GPU and CUDA platform of general purpose computing is evaluated in this study. Its architecture and memory architecture is unique. In addition, Mathematica is used in this study for file IO and graphic output with unique CUDA interface.
Then, the each ray tracing method is validated using a spherical lens, aspherical lens, and a headlamp lens. From the comparison, the double precision floating Nagata triangular patch method is best in accuracy. Acceleration of ray tracing using CUDA was successful having 2 times implement in 362 million rays traced, compared to commercially available ray trace packages under the same computing resources.
| Identifer | oai:union.ndltd.org:arizona.edu/oai:arizona.openrepository.com:10150/626712 |
| Date | January 2017 |
| Creators | Kimura, Ryota, Kimura, Ryota |
| Contributors | Chipman, Russell A., Chipman, Russell A., Koshel, Richard J., Takashima, Yuzuru |
| Publisher | The University of Arizona. |
| Source Sets | University of Arizona |
| Language | en_US |
| Detected Language | English |
| Type | text, Electronic Thesis |
| Rights | Copyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author. |
Page generated in 0.0025 seconds