Low-discrepancy Sequence Path Tracing

发布时间：2024-02-25 22:28

　　计算机图形学是计算机科学的一门重要的分支,它研究的主要内容包含图形在计算机中的表示、图形的生成与处理和图形的显示。计算机图形学广泛用于电影特效,游戏制作,虚拟现实和增强现实等领域。在计算机绘制图像中,首先需要使用特定的数据结构表示计算机建立的模型,然后再将模型转换成屏幕上显示的图像;这个转换过程,就称为渲染,有些时候也叫做着色。计算机中相同的模型使用不同的渲染方式可以生成不同风格的图像;在不同的应用场景中,比如不同风格的电影,又或者是实时和非实时渲染;这些不同的场景中,采用的渲染技术是有所差别的。所以渲染是计算机图形学中重要的研究领域,其中真实感渲染更是广大研究者的研究重点。真实感渲染技术要求按照现实生活中的场景和光影生成图像。尽管在实际应用中,很多渲染技术都是采用了一些近似方法来模拟真实的光影表现,但是真实感渲染技术的最终目的都是生成照片级的图像。生成真实感图像主要包含两个关键步骤,第一步是精确的场景模型表示。由于计算机无法使用连续的值来对场景进行描述,所以,首先需要用离散的值精确对场景进行描述。第二步则需要使用真实感的渲染技术。早期,由于计算机硬件的限制,渲染技术一直停留在光栅化流...

【文章页数】：77 页

【学位级别】：硕士

【文章目录】：
Abstract
1 Introduction
    1.1 Research Background and Significance
    1.2 Research Actuality
        1.2.1 Graphic Processing Algorithms
        1.2.2 The Development and Status of Hardware Architecture of GPU
    1.3 Research Content
    1.4 Chapter Arrangement
2 Related Concepts
    2.1 Research Motivation
    2.2 Geometric Optical Model
        2.2.1 Illumination Measurement
        2.2.2 Bidirectional Reflectance Distribution Function (BRDF)
    2.3 Rendering Equation and Ray Tracing
    2.4 Path Tracing
    2.5 Monte-Carlo Integral
3 Monte-Carlo Path Tracing and Implementation of Common Illumination Models
    3.1 Research Motivation
    3.2 Overview
    3.3 Monte-Carlo Path Tracing Implementation and Convergence Analysis
        3.3.1 Algorithms Overview
        3.3.2 Intersection Calculation
        3.3.3 Calculation the Color of Pixel
        3.3.4 Depth Control
        3.3.5 Experimental Result and Instance Display
    3.4 Experiment on Convergence Analysis of Termination Conditions
    3.5 Implementation of Illumination Models
        3.5.1 Overview of Illumination Models
        3.5.2 Specular Reflection
        3.5.3 Diffuse Reflection
    3.6 Experimental Result and Analysis
4 Low-discrepancy Ray Tracing
    4.1 Research Motivation
    4.2 Metrics of the Global Illumination Algorithm
        4.2.1 Robustness, Rendering Speed and Accuracy
        4.2.2 Discrepancy and Consistency
    4.3 Low-Discrepancy Sequence Path Tracing
        4.3.1 Convergence analysis of Path Tracing
    4.4 Experiment Result and Analysis
        4.4.1 Comparison of Convergence between Low Discrepancy Sequencesand Random Sequences
        4.4.2 Efficient Implementation of Low-Discrepancy Sequences
        4.4.3 Unbiased Low-Discrepancy Sequence Path Tracing Algorithms
        4.4.4 Conclusion
5 Parallel Low-Discrepancy Sequence Path Tracing Algorithm
    5.1 Research Motivation
    5.2 Parallel Low-Discrepancy Sequence Path Tracing Algorithm
        5.2.1 Parallel Low-Discrepancy Sequence Path Tracing Based on OpenMP
        5.2.2 Parallel Low-Discrepancy Sequence Path Tracing Based on CUDA
    5.3 Experimental Results and Analysis
6 Conclusion
References
Acknowledgements
Appendix A 中文摘要



本文编号：3910940