裸眼3D显示关键技术研究.pdf

发布时间：2022-06-01 发布人：admin 分类：说明书资料大小：5.69M 资料格式：pdf 举报版权申诉

kolaking-12745198-4744300845385551485.pdf-第1页.png

第1页 / 共114页

kolaking-12745198-4744300845385551485.pdf-第2页.png

第2页 / 共114页

kolaking-12745198-4744300845385551485.pdf-第3页.png

第3页 / 共114页

kolaking-12745198-4744300845385551485.pdf-第4页.png

第4页 / 共114页

kolaking-12745198-4744300845385551485.pdf-第5页.png

第5页 / 共114页

kolaking-12745198-4744300845385551485.pdf-第6页.png

第6页 / 共114页

kolaking-12745198-4744300845385551485.pdf-第7页.png

第7页 / 共114页

kolaking-12745198-4744300845385551485.pdf-第8页.png

第8页 / 共114页

摘要

ABSTRACT

图表目录

术语表

1 绪论

1.1 立项依据

1.1.1 研究意义

1.1.2 国内外研究现状

1.2 研究内容

2 裸眼3D显示原理

引言

2.1 裸眼如何感知3D

2.2 3D视频源格式

2.2.1 左右视图

2.2.2 上下视图

2.2.3 视频+深度图(video plus depth, V+D)

2.2.4 多路视频+深度(multi-view video plus depth, MVD)

2.2.5 分层深度视频(layered depth video, LDV)

2.3 裸眼3D显示技术

2.3.1 光屏障式技术

2.3.2 柱状透镜技术

2.3.3 集成成像技术

2.4 本章小结

3 裸眼3D内容制作

引言

3.1 3D内容制作技术简介

3.1.1 3D拍摄

3.1.2 2D转3D技术

3.1.3 计算机3D模型绘制

3.1.4 视图合成

3.2 基于Autodesk® Maya软件的裸眼3D内容制作

3.3 基于DIBR的裸眼3D内容生成算法

3.3.1 无需摄像机参数的三维图像变换函数

3.3.2 基于视差图中值滤波的匹配误差消除算法

3.3.3 基于视图判定的褶皱消除方法

3.3.4 基于视差图的空洞填充算法

3.3.5 算法仿真与结果

3.4 裸眼3D内容的合成

3.5 本章小结

4 裸眼3D内容播放

引言

4.1 裸眼3D内容播放器的系统框图

4.2 DIBR系统的硬件实现

4.2.1 系统架构

4.2.2 行级流水线

4.3 多视图图像融合算法及实现

4.3.1 多视图图像融合算法

4.3.2 算法仿真与结果

4.3.3 8视点图像融合算法的硬件实现

4.4 本章小结

5 硬件开发平台及实验结果分析

5.1 硬件开发平台及开发工具

5.2 测试方案设计

5.3 实验结果展示与分析

5.3.1 DIBR系统实验结果及合成性能分析

5.3.2 图像融合系统实验结果及分析

5.3.3 实验结果上屏展示及分析

5.4 本章小结

6 总结与展望

6.1 总结

6.2 展望

致谢

参考文献

附录

A. 作者攻读期间发表的相关科研论文

B. 作者攻读期间申请的相关专利

C. 作者攻读期间参与的相关科研项目

D. 相关代码的SVN路径

裸眼 3D 显示关键技术研究重庆大学硕士学位论文（学术学位）学生姓名：谭伟敏指导教师：黄扬帆教授刘然副教授专业：电路与系统学科门类：工学重庆大学通信工程学院二 O 一四年四月

Research on Key Technologies of Glasses-Free 3D Display A Thesis Submitted to Chongqing University in Partial Fulfillment of the Requirement for the Master’s Degree of Engineering By Tan Weimin Supervisor : Prof. Huang Yangfan Assistant Supervisor: Associate Prof. Liu Ran Specialty: Circuit and System College of Communication Engineering of Chongqing University, Chongqing, China April, 2014

中文摘要摘要近年来，随着 3D 显示技术的快速发展，3D 产品变得越来越流行。然而，目前大多数的普通 3D 产品仍然需要观众佩戴特殊的眼镜，这严重制约了 3D 产品的普及。为解决这一问题，裸眼 3D 显示技术应运而生，并受到广泛关注。本论文的研究目的是通过对裸眼 3D 显示的关键技术进行研究，解决裸眼 3D 显示面临的内容制作和播放问题，为裸眼 3D 显示的应用提供理论和技术支撑。论文首先介绍了裸眼 3D 显示原理，然后在该原理的基础上，对裸眼 3D 显示的关键技术——裸眼 3D 内容制作和裸眼 3D 内容播放展开研究。在裸眼 3D 内容制作方面，重点研究了基于 Autodesk® Maya 软件的裸眼 3D 内容制作方法和基于深度图像绘制(depth-image-based rendering, DIBR)的裸眼 3D 内容生成算法；在裸眼 3D 内容播放方面，提出了裸眼 3D 播放器的硬件架构，并给出了基于 DIBR 算法以及多视图图像融合算法的硬件实现。论文意在获取高质量裸眼 3D 内容和裸眼 3D 播放效果的同时，降低算法的复杂性，以便硬件实现。论文的主要研究成果包括： ① 根据shift-sensor摄像机设置的性质，分析并改进了基于DIBR的视图合成算法，并研究了采用改进算法合成多视点视图的原理。改进后的算法和原算法相比，合成新视点的图像质量相当，适合硬件实现。 ② 介绍了基于 Autodesk® Maya软件制作裸眼3D 内容的方法。该方法采用 CelVision 3D Creator插件，通过设置参数、调整不同平面之间的距离，绘制得到8 幅具有视差的图像。该方法虽然复杂、耗时较长，但制作过程灵活，制作的裸眼 3D内容质量更高、3D效果更明显。 ③ 提出了一种多视图图像融合算法。该算法首先对含有人眼视差的多幅视图进行子像素抽取，然后按照光栅子像素排列规律进行排列，最后合成得到适合基于光栅的裸眼立体显示器显示的合成图像。实验表明该算法可对8幅全高清视图进行有效的融合，融合得到的合成图像适合于基于光栅的裸眼立体显示器显示，且算法硬件实现简单。 ④ 在FPGA开发板上，用较少的硬件开销实现了DIBR系统和图像融合系统。相应地，给出了硬件架构、工作原理、时序图和仿真结果。本论文力求在研究方法与思路上有所突破，其研究成果将为裸眼 3D 显示的应用提供具有实用价值的新方法。关键词：3D 显示，裸眼，图像融合，DIBR，FPGA I

重庆大学硕士学位论文 II

英文摘要 ABSTRACT With the rapid development of 3D display technology in recent years, 3D products have become more and more popular. However, the most common 3D products require users to wear special glasses, which has seriously limited the popularity of 3D products. To solve this problem, glasses-free 3D display technology comes into being and has been widely concerned. The purpose of this dissertation is to solve the problems of content creation and play facing glasses-free 3D display by researching on key technologies of glasses-free 3D display, which will provide the theoretical and technical support for the application of glasses-free 3D display. In this dissertation, the theoretical basis of glasses-free 3D display is firstly introduced. Then, based on the theoretical basis, the key techniques of glasses-free 3D display, glasses-free 3D content creation and play, are studied. As for content creation, two glasses-free 3D content creation methods, the method based on Autodesk® Maya software and the method based on depth-image-based rendering (DIBR), are thoroughly studied; as for content play, a hardware architecture of glasses-free 3D player is proposed. Moreover, the hardware implementation of DIBR and multi-view image fusion algorithm is deliberated. This dissertation aims at achieving high quality of glasses-free 3D contents and player effects with relatively low complexity of algorithms which can facilitate hardware implementation. The main research results of this dissertation are summarized as follows. ① According to the characteristics of shift-sensor camera setup, the view synthesis algorithm based on DIBR is analyzed and improved, and the theory of multi-view synthesis based on the improved algorithm is studied. Compared with original algorithm, the improved algorithm can provide comparable image quality, while it is facilitative for hardware implementation. ② A glasses-free 3D contents creation method based on Autodesk® Maya software is introduced. This method can render 8 views using CelVision 3D Creator plugin by setting parameters to adjust the distance between different planes. Although the method is complex and time-consuming, it is flexible for content creation. In addition, the quality of created glasses-free 3D contents is higher, and 3D effects are better. III

重庆大学硕士学位论文 ③ A multi-view image fusion algorithm is proposed. Firstly, the algorithm extracts sub-pixels from multi views which contain parallaxes, and then arrange these sub-pixels in accordance with the sub-pixel arrangement rule for parallax barrier. Finally, composite image is obtained which adapts to glasses-free 3D displays based on parallax barrier. Experimental results show that the proposed approach can fuse 8 Full High Definition (FHD) views effectively, and the composite image can be displayed on glasses-free 3D displays based on parallax barrier. Moreover, this algorithm is easy for hardware implementation. ④ The DIBR system and image composite system are implemented on a development platform of FPGA with low hardware cost. Correspondingly, the hardware architecture, operational principle, timing diagram and simulation results are also presented. As an original work, the author tried his best to make a breakthrough on research methods and research thoughts, and the research results will provide new methods for the practical applications of glasses-free 3D display. Keywords: 3D display, Glasses-free, image fusion, DIBR, FPGA IV

分享到：

赞收藏

资料库

裸眼3D显示关键技术研究.pdf

相关推荐

行业

热门标签

最新资料