Project Abstract

One of the main goals of the computer graphics industry has been the ability to create photo-realistic renders of digital scenes. Recently, ray tracing has become the standard approach towards processing these computer-generated scenes that require photorealistic lighting effects. This dissertation project was to implement a high-quality offline path tracing software that implements both hardware and software acceleration techniques to create realistic visuals with improved rendering times. This was completed by implementing the main rendering algorithm in CUDA, allowing the software to use NVIDIA GPU hardware to render individual image pixels in parallel, and by implementing the KD-Tree, a space-partitioning data structure to reduce the search area of each individual ray. These improvements allowed for the addition of more complex features, such as support for .OBJ 3D models and UV texture mapping. The final project allows us to render complex scenes of up to 2 million polygons within a respectable timeframe.

Keywords: Computer graphics, Raytracing, GPU programming

 Conference Details

Session: B

Location: Sir Stanley Clarke Auditorium at 13:30 15:30

Markers: Mark Jones, Simeng Qiu

Course: BSc Computer Science 3yr FT

Future Plans: I’m looking for work