基于GPU加速的光线跟踪体绘制算法研究
|
摘要
早期,光线跟踪体绘制算法主要是靠CPU来实现,而GPU主要是用于图形处理。本文将利用GPU强大的并行处理能力和可编程性,将GPU视为流处理器来实现算法的加速处理工作。
首先,针对光线跟踪体绘制算法经常用到的均匀网格,KD-tree,包围体层次这三个加速结构在GPU上进行创建和遍历的细节问题进行了阐述和分析研究。
其次,详细对GPU的图形渲染管线技术和流处理开发工具进行了介绍,并且通过具体实现一个流处理模型的设计,来详细具体介绍如何将基于GPU的光线跟踪体绘制算法实现的。
最后,通过在不同场景复杂度和不同分辨率等情况下做对比试验。通过对比说明,更能体现GPU的优越性,并且通过对比来实现算法的进一步的改进。
Early stage, ray tracing algorithm for volume rendering is mainly achieved by the CPU, and the GPU is mainly used for graphics processing. In the paper, we will take advantage of the GPU parallel processing capability and powerful programmable, and take the GPU as a stream processor to accelerate the processing algorithm.
Firstly, we have elaborate the details about create and traverse based on GPU of the three acceleration structures like Uniform Grid, KD-tree, BVH, which are often used by ray tracing algorithm for volume rendering, the paper compare and analyze the three acceleration structures.
Secondly, we have a detailed introduce for the graphics rendering pipeline of the GPU and development tools for stream processing, and through the design of a stream processing model, ray tracing for volume-rendering based on GPU has been detail introduced how to make it realized.
Lastly, we have a comparison tests under different complicated scene and resolutions. By comparison shows that better reflect the advantages of GPU, and by comparing the algorithm to achieve further improvements.
首先,针对光线跟踪体绘制算法经常用到的均匀网格,KD-tree,包围体层次这三个加速结构在GPU上进行创建和遍历的细节问题进行了阐述和分析研究。
其次,详细对GPU的图形渲染管线技术和流处理开发工具进行了介绍,并且通过具体实现一个流处理模型的设计,来详细具体介绍如何将基于GPU的光线跟踪体绘制算法实现的。
最后,通过在不同场景复杂度和不同分辨率等情况下做对比试验。通过对比说明,更能体现GPU的优越性,并且通过对比来实现算法的进一步的改进。
Early stage, ray tracing algorithm for volume rendering is mainly achieved by the CPU, and the GPU is mainly used for graphics processing. In the paper, we will take advantage of the GPU parallel processing capability and powerful programmable, and take the GPU as a stream processor to accelerate the processing algorithm.
Firstly, we have elaborate the details about create and traverse based on GPU of the three acceleration structures like Uniform Grid, KD-tree, BVH, which are often used by ray tracing algorithm for volume rendering, the paper compare and analyze the three acceleration structures.
Secondly, we have a detailed introduce for the graphics rendering pipeline of the GPU and development tools for stream processing, and through the design of a stream processing model, ray tracing for volume-rendering based on GPU has been detail introduced how to make it realized.
Lastly, we have a comparison tests under different complicated scene and resolutions. By comparison shows that better reflect the advantages of GPU, and by comparing the algorithm to achieve further improvements.
引文
[1]McCormick BH, DeFanti T A, Brown M D. Visualization in scientific Computing. Computer Graphics,1987,1(2):99-108
[2]店洋圣 等,三维数据场可视化,北京：清华大学出版社社,1999
[3]Lorensen W E, Cline H E, Marching Cube:a High-resolution 3D Surface Construction Algorithm. Computer Graphics,1987.21 (4):163-169
[4]葛军,舒华忠,鲍旭东基于面绘制的GPU体绘制加速应用科学学报2009,27(2)：167-171
[5]费耀平 胡慧君,陈松乔等基于GPU图像直接体绘制算法分析与评价计算机应用研究2009,26(6)：2020-2022
[6]John Montrym. The GeForce 6800. IEEE Computer Society,2005,25(2):41-51
[7]彭延军,石教英,体绘制技术在医学可视化中的新发展,中国图像钊形学报,2002,12(2)：10-16
[8]Sungup C, Hngeongdo K. Efficient Space Leaping Method for Volume Rendering. Medical Imaging 1999, SHE Conference on Visual Data Exploration and Analysis. Munich:SPIE Press,1999,263-270.
[9]Levoy M. Volume Rendering by Adaptive. Visual Computer.1990,6(1);2-7
[10]T Gunther, C Poliwoda, C Reinhart, J Hesser. A Massively Parallel Processor for Realtime Volume Visualization in Medicine. Proc. Euro graphics Workshop, Oslo, Norway,1994:103-108
[11]夏放怀,沈振康等,一种用于实时体绘制系统的自适应采样算法,电子学报,2002.3
[12]Kruger. J, R. Westermann. Acceleration Techniques for GPU-based Volume Rendering. Proceedings of the 14th IEEE Visualization 2003 (VIS'03) 2003
[13]Repplinger M, Loffler A, Thielen M, et al. A Flexible Adaptation Service for Distributed Rendering. In:Proceedings of the 9th Eurographics Symposium on Parallel Graphics and Visualization. Munich, Germany,2009,29-30
[14]Gregory J W, Francis M R, Robert D C. A ray tracing solution for diffuse interreflection. In: Proceedings of International Conference on Computer Graphics and Interactive Techniques. San Diego, California, USA,2007,85-92
[15]Pabst H F, Springer J P, Schollmeyer A, Lenhardt R, et al. Ray casting of trimmed NURBS surfaces on the GPU. In:Proceedings of the 2006 IEEE Symposium on Interactive Ray Tracing. Salt Lake City, Utah, USA,2006,151-160
[16]GroB M, Lojewski C, Bertram M, et al. Fast Implicit KD-Trees:Accelerated Isosurface Ray Tracing and Maximum Intensity Projection for Large Scalar Fields. In Proceedings of Computer Graphics and Imaging. Innsbruck, Austria,2007,67-74
[17]Stegmaier S. A simple and flexible volume renderinrendering framework for graphics-hardware-based raycasting[C]. Proceedings of the International Workshop on Volume Graphics,2005:187-195
[18]Ullman T, Preidel T, Bruderlin B. Effcient Sampling of Textured Scenes in Vertex Tracing. In: Proceedings of Eurographics Conference. Granada, Spain,2003,126-133
[19]黄沛杰,王文成,杨刚等.基于中介面加快光线跟踪计算.计算机学报,2007,30(2)：262-271
[20]Woop S, Marmitt G, Slusallek P. B-KD trees for hardware accelerated ray tracing of dynamic scenes. In:Proceedings of the 21st ACM SIGGRAPH/Eurographics symposium on Graphics hardware. Vienna, Austria, 2006,67-77
[21]苏超轼,赵明昌,张向文GPU加速的八叉树体绘制算法计算机应用2008,28(5)：1232-1239
[22]邹华,高新波,吕新荣.一种八叉树编码加速的3D纹理体绘制算法.西安交通大学学报,2008,42(12)：1490-1494
[23]杨俊华,符红光,郭惠 基于GPU快速光线跟踪算法的设计与实现计算机应用2007,27(8)：2032-2035
[24]Zhou K, Hou Q, Wang R, et al. Real-time KD-tree construction on graphics hard-ware. ACM Transactions on Graphics,2008,27(5):1-11
[25]Matt Pharr and Greg Humphreys. Physically Based Rendering. Morgan Kauffman,2004
[26]Eric Haines. BSP plane cost function revisited. Ray Tracing News,2006,17(1):222-228
[27]邹华 高新波 吕新荣 层次包围盒与GPU实现相结合的光线投射算法计算机辅助设计与图形学学报2009,21(2)：172-178
[28]Jeffrey Goldsmith and John Salmon. Automatic creation of object hierarchies for ray tracing. IEEE Computer Graphics and Applications,1987,7:14-20
[29]Fatahalian K, Houston M. A Closer Look at GPUS. Communications of the ACM,2008,51(10),566-574
[30]Owens J, Houston M, Luebke D, et al. GPU Computing. In:Proceedings of the IEEE. Piscataway, New Jersey, USA:IEEE,2008,879-899
[31]Buck I, Foley T, Horn D, et al. Brook for GPUs:stream computing on graphics hardware. ACM Transactions on Graphics,2004,23(3):777-786
[32]郭际GPU才是并行运算的未来中国计算机报2008,4(14)：13-17
[33]刘姚新基十GPU的实时绘制算法研究：[硕士学位论文]重庆：计算机软件与理论,2007
[34]Purcell J T. Ray Tracing on a Stream Processor:[PhD thesis]. Stanford, California, USA:Computer Graphics Lab of Stanford University,2004,15-226
[35]Christen M. Ray Tracing on GPU:[Master's thesis]. Switzerland:Fachhochschule beider Basel,2005,12-36
[36]王碧薇 基于GPU的光线跟踪算法的分析科技资讯2007,23(6)：5-6
[37]章圣洁,王国荣,王小平 基于GPU加速的光线跟踪技术研究Computer Era 2007,6:36-38
[38]Mailer T, Trumbore B. Fast, Minimum Storage Ray-Triangle Intersection. Journal of Graphics Tools, 1997,2(1):21-28
[39]张延红,周必水 基于GPU加速的光线与三角面片求交Computer Era 2006,5:29-31
[40]李建明,万单领,何荣盛等大连理工大学学报2008，48(2)：287-291
[2]店洋圣 等,三维数据场可视化,北京：清华大学出版社社,1999
[3]Lorensen W E, Cline H E, Marching Cube:a High-resolution 3D Surface Construction Algorithm. Computer Graphics,1987.21 (4):163-169
[4]葛军,舒华忠,鲍旭东基于面绘制的GPU体绘制加速应用科学学报2009,27(2)：167-171
[5]费耀平 胡慧君,陈松乔等基于GPU图像直接体绘制算法分析与评价计算机应用研究2009,26(6)：2020-2022
[6]John Montrym. The GeForce 6800. IEEE Computer Society,2005,25(2):41-51
[7]彭延军,石教英,体绘制技术在医学可视化中的新发展,中国图像钊形学报,2002,12(2)：10-16
[8]Sungup C, Hngeongdo K. Efficient Space Leaping Method for Volume Rendering. Medical Imaging 1999, SHE Conference on Visual Data Exploration and Analysis. Munich:SPIE Press,1999,263-270.
[9]Levoy M. Volume Rendering by Adaptive. Visual Computer.1990,6(1);2-7
[10]T Gunther, C Poliwoda, C Reinhart, J Hesser. A Massively Parallel Processor for Realtime Volume Visualization in Medicine. Proc. Euro graphics Workshop, Oslo, Norway,1994:103-108
[11]夏放怀,沈振康等,一种用于实时体绘制系统的自适应采样算法,电子学报,2002.3
[12]Kruger. J, R. Westermann. Acceleration Techniques for GPU-based Volume Rendering. Proceedings of the 14th IEEE Visualization 2003 (VIS'03) 2003
[13]Repplinger M, Loffler A, Thielen M, et al. A Flexible Adaptation Service for Distributed Rendering. In:Proceedings of the 9th Eurographics Symposium on Parallel Graphics and Visualization. Munich, Germany,2009,29-30
[14]Gregory J W, Francis M R, Robert D C. A ray tracing solution for diffuse interreflection. In: Proceedings of International Conference on Computer Graphics and Interactive Techniques. San Diego, California, USA,2007,85-92
[15]Pabst H F, Springer J P, Schollmeyer A, Lenhardt R, et al. Ray casting of trimmed NURBS surfaces on the GPU. In:Proceedings of the 2006 IEEE Symposium on Interactive Ray Tracing. Salt Lake City, Utah, USA,2006,151-160
[16]GroB M, Lojewski C, Bertram M, et al. Fast Implicit KD-Trees:Accelerated Isosurface Ray Tracing and Maximum Intensity Projection for Large Scalar Fields. In Proceedings of Computer Graphics and Imaging. Innsbruck, Austria,2007,67-74
[17]Stegmaier S. A simple and flexible volume renderinrendering framework for graphics-hardware-based raycasting[C]. Proceedings of the International Workshop on Volume Graphics,2005:187-195
[18]Ullman T, Preidel T, Bruderlin B. Effcient Sampling of Textured Scenes in Vertex Tracing. In: Proceedings of Eurographics Conference. Granada, Spain,2003,126-133
[19]黄沛杰,王文成,杨刚等.基于中介面加快光线跟踪计算.计算机学报,2007,30(2)：262-271
[20]Woop S, Marmitt G, Slusallek P. B-KD trees for hardware accelerated ray tracing of dynamic scenes. In:Proceedings of the 21st ACM SIGGRAPH/Eurographics symposium on Graphics hardware. Vienna, Austria, 2006,67-77
[21]苏超轼,赵明昌,张向文GPU加速的八叉树体绘制算法计算机应用2008,28(5)：1232-1239
[22]邹华,高新波,吕新荣.一种八叉树编码加速的3D纹理体绘制算法.西安交通大学学报,2008,42(12)：1490-1494
[23]杨俊华,符红光,郭惠 基于GPU快速光线跟踪算法的设计与实现计算机应用2007,27(8)：2032-2035
[24]Zhou K, Hou Q, Wang R, et al. Real-time KD-tree construction on graphics hard-ware. ACM Transactions on Graphics,2008,27(5):1-11
[25]Matt Pharr and Greg Humphreys. Physically Based Rendering. Morgan Kauffman,2004
[26]Eric Haines. BSP plane cost function revisited. Ray Tracing News,2006,17(1):222-228
[27]邹华 高新波 吕新荣 层次包围盒与GPU实现相结合的光线投射算法计算机辅助设计与图形学学报2009,21(2)：172-178
[28]Jeffrey Goldsmith and John Salmon. Automatic creation of object hierarchies for ray tracing. IEEE Computer Graphics and Applications,1987,7:14-20
[29]Fatahalian K, Houston M. A Closer Look at GPUS. Communications of the ACM,2008,51(10),566-574
[30]Owens J, Houston M, Luebke D, et al. GPU Computing. In:Proceedings of the IEEE. Piscataway, New Jersey, USA:IEEE,2008,879-899
[31]Buck I, Foley T, Horn D, et al. Brook for GPUs:stream computing on graphics hardware. ACM Transactions on Graphics,2004,23(3):777-786
[32]郭际GPU才是并行运算的未来中国计算机报2008,4(14)：13-17
[33]刘姚新基十GPU的实时绘制算法研究：[硕士学位论文]重庆：计算机软件与理论,2007
[34]Purcell J T. Ray Tracing on a Stream Processor:[PhD thesis]. Stanford, California, USA:Computer Graphics Lab of Stanford University,2004,15-226
[35]Christen M. Ray Tracing on GPU:[Master's thesis]. Switzerland:Fachhochschule beider Basel,2005,12-36
[36]王碧薇 基于GPU的光线跟踪算法的分析科技资讯2007,23(6)：5-6
[37]章圣洁,王国荣,王小平 基于GPU加速的光线跟踪技术研究Computer Era 2007,6:36-38
[38]Mailer T, Trumbore B. Fast, Minimum Storage Ray-Triangle Intersection. Journal of Graphics Tools, 1997,2(1):21-28
[39]张延红,周必水 基于GPU加速的光线与三角面片求交Computer Era 2006,5:29-31
[40]李建明,万单领,何荣盛等大连理工大学学报2008，48(2)：287-291