当前位置:主页 > 文艺论文 > 动漫艺术论文 >

异构建模渲染系统数据转换方法研究

发布时间:2018-06-04 18:48

  本文选题:数据转换 + 建模软件 ; 参考:《山东大学》2009年硕士论文


【摘要】: 在图形制作中,渲染是3D创意的基础,也是3D制作中决定最终图像效果的关键环节。目前商业、非商业的渲染器广泛应用于动漫制作、科学模拟、图形绘制等多个领域.着CG技术的发展,渲染的计算量日益增加,单个计算机难以满足巨大的计算量需求,集群渲染应运而生。集群渲染平台可以提供公共渲染服务.但在实际应用中,用户使用不同建模软件制作场景,选择某种渲染器来实现某种渲染特效,这要求渲染平台安装相应的渲染器,以满足提供公共渲染服务的需求。 CG技术发展过程中一直没有制定统一的工业标准,导致目前广泛使用的建模软件(如Maya、3DMax)、渲染软件(如mental ray、RenderMan等)所采用的数据结构、数据组织方法和数据在系统中的存储和表现形式都各不相同。由于渲染器的输入接口不统一,导致模型的制作过程必须与渲染器紧密联系在一起,很多时候需要手工转化数据。目前的3D建模软件模型转换功能并不完美,经常会出现数据元素缺失的情况,造成场景严重失真。异构的建模软件及渲染器之间的模型转换问题,对软件间集成应用造成了障碍,因此如何共享建模、渲染软件已有的模型资源,处理异构系统的模型之间的转换,已经成为系统开发中迫切需要解决的问题。 针对当前集群渲染平台的需求,本文以RenderMan和mental ray的接口标准为主要设计依据,对异构渲染系统数据转换进行了研究,提出了一种基于中性文件的异构数据模型转换的模板。根据3D数据模型的特点,提出模型转换的映射方法,经过一系列的特征抽取、属性匹配、应用封装将建模软件的模型转换为渲染软件的模型。进一步,给出以该转换方法为指导的实现工具:MayaExport,通过实例证明该方法的有效性。 本文的主要工作包括: 1.针对渲染器模型描述标准不统一的问题,以RenderMan、mental ray标准为依据,采用分层思想,制定可扩展的模型转换中性文件。渲染器通过插件或修改源码的方式增加新功能时,中性文件通过添加新标签可以提供支持. 2.针对建模软件与渲染器的异构问题,在Maya、3DMax等建模软件及mental ray、RenderMan等渲染器标准的基础上,提出基于数据和命令的模型转换方法。 3.针对渲染器内部算法差异造成的转换过程模型精度差异与数值理解误差问题,通过集成建模软件与渲染器,交互的控制转换的等级和参数的调节。 本文的工作已经成功地用于山东省动漫渲染平台中。 在上述研究的基础上,本文还需要在着色器和过程纹理等方面进一步开展工作。
[Abstract]:Rendering is not only the foundation of 3D creation, but also the key link to determine the final image effect. At present, commercial, non-commercial renderer is widely used in animation production, scientific simulation, drawing and other fields. With the development of CG technology, the computation of rendering is increasing day by day, so it is difficult for a single computer to meet the huge demand of computation, so cluster rendering emerges as the times require. The cluster rendering platform can provide public rendering services. However, in practical applications, users use different modeling software to create scenes and select a renderer to implement a certain rendering effect, which requires the rendering platform to install the corresponding renderer to meet the requirements of providing public rendering services. During the development of CG technology, there has been no uniform industrial standard, which leads to the data structure used in the widely used modeling software (such as Maya3 D MaxX, rendering software (such as mental rayne RenderMan, etc.). Data organization methods and data storage and representation in the system are different. Because the input interface of the renderer is not uniform, the process of making the model must be closely related to the renderer. In many cases, it is necessary to transform the data manually. The current 3D modeling software model transformation function is not perfect, often there are data elements missing, resulting in serious distortion of the scene. The problem of model conversion between heterogeneous modeling software and renderer creates obstacles to the application of software integration, so how to share modeling, render the existing model resources of the software, and handle the transformation between models of heterogeneous systems. It has become an urgent problem in system development. According to the demand of the current cluster rendering platform, based on the interface standard of RenderMan and mental ray, this paper studies the data conversion of heterogeneous rendering system, and proposes a template for the conversion of heterogeneous data model based on neutral files. According to the characteristics of 3D data model, the mapping method of model transformation is proposed. After a series of feature extraction and attribute matching, the model of modeling software is transformed into the model of rendering software by encapsulation. Furthermore, a realization tool named: Maya Export, which is guided by the transformation method, is given, and the effectiveness of the method is proved by an example. The main work of this paper includes: 1. In order to solve the problem that the model description standard of renderer is not uniform, based on the Render mental ray standard, an extensible model conversion neutral file is developed by using the idea of stratification. When the renderer adds new functionality by plug-in or by modifying the source code, neutral files can be supported by adding new tags. 2. Aiming at the isomerism of modeling software and renderer, a model transformation method based on data and command is proposed based on the modeling software such as Maya3DMax and mental renderman. 3. In order to solve the problem of the model accuracy difference and numerical understanding error caused by the algorithm difference within the renderer, the level and parameter adjustment of the conversion are interactively controlled by integrating the modeling software with the renderer. The work of this paper has been successfully used in the animation rendering platform of Shandong Province. On the basis of the above research, this paper needs further research on shader and process texture.
【学位授予单位】:山东大学
【学位级别】:硕士
【学位授予年份】:2009
【分类号】:TP391.41

【引证文献】

相关硕士学位论文 前1条

1 甄菲菲;渲染引擎研发中关键技术的研究与实现[D];山东大学;2010年



本文编号:1978469

资料下载
论文发表

本文链接:https://www.wllwen.com/wenyilunwen/dongmansheji/1978469.html


Copyright(c)文论论文网All Rights Reserved | 网站地图 |

版权申明:资料由用户dbaff***提供,本站仅收录摘要或目录,作者需要删除请E-mail邮箱bigeng88@qq.com