基于人体动态采集的三维建模与交互

 2022-05-16 08:05

论文总字数:26954字

摘 要

三维建模技术将现实世界的物体转换为数字化的三维模型,并利用计算机进行显示和操作。静态的三维建模已经无法满足人们日益增长的需求,因此专家学者们致力于研究动态三维建模技术。动态三维建模技术的研究重点,一是如何逼真地复现目标对象,二是如何加强目标对象和虚拟模型的交互。研究成果能被广泛应用于3D电影特效制作、虚拟试衣系统、体感游戏、模拟解剖软件等多个领域。传统的三维扫描仪器体积庞大、价格昂贵、操作复杂,但深度摄像头的出现解决了这些问题,它们便携小巧、性价比高。本文提出了一种基于RealSense D415摄像头和Unity 3D驱动的三维人体模型交互系统,主要做了以下工作:

首先,总结相关的技术方法和原理,比如三维人体建模技术、动态捕获技术,调研国内外研究现状。

其次,介绍实验项目中涉及的RealSense技术、D415摄像头的工作原理和关键参数,以及Nuitrack SDK所提供的骨骼追踪技术。

然后,本课题通过RealSense D415来捕获深度信息,并为后续的三维建模进行相应的处理操作。

最后,在Unity 3D平台中,利用Nuitrack SDK实现目标对象和虚拟模型的交互。

关键词:RealSense D415;Unity 3D;Nuitrack;骨骼追踪;动态三维建模与交互

ABSTRACT

3D modeling technology transforms real-world objects into digital 3D models, displays and manipulates them by using a computer. Static 3D modeling has been unable to meet the growing needs of people, so experts and scholars are committed to research dynamic 3D modeling technology. The first research focus of dynamic 3D modeling technology is how to realistically reproduce the target object, and the second focus is how to strengthen the interaction between the target object and the virtual model. The research results can be widely used in 3D film special effects production, virtual fitting system, somatosensory game, simulation anatomy software and other fields. Traditional 3D scanning instruments are bulky, expensive, and complicated to operate, but the emergence of depth cameras solves these problems, and they are compact and cost-effective. This paper proposes a 3D human body model interaction system based on RealSense D415 camera and Unity 3D driver, which mainly does the following work:

First, summarize the relevant technical methods and principles, such as 3D human body modeling technology and dynamic capture technology, and investigate the research status at home and abroad.

Secondly, it introduces the RealSense technology involved in the experimental project, the working principle and key parameters of the D415 camera, and the skeleton tracking technology provided by the Nuitrack SDK.

Then, the project captures depth information by using RealSense D415, and performs corresponding processing operations for subsequent 3D modeling.

Finally, in the Unity 3D platform, the interaction between the target object and the virtual model is implemented using the Nuitrack SDK.

Key Words: RealSense D415; Unity 3D; Nuitrack; Skeleton Tracking; Dynamic 3D Human Body Modeling and Interaction

目 录

第一章 绪论 1

1.1三维建模的研究背景及意义 1

1.2国内外研究现状 2

1.3研究内容 3

1.4论文章节安排 3

第二章 动态三维建模的相关技术 5

2.1三维人体建模方法 5

2.1.1基于三维建模软件的三维人体建模 5

2.1.2基于人体扫描技术的三维人体建模 5

2.1.3基于人体图像的三维人体建模 6

2.2动态捕捉技术 7

2.2.1机械式动态捕捉 7

2.2.2声学式动态捕捉 7

2.2.3电磁式动态捕捉 8

2.2.4惯性动态捕捉 8

2.2.5光学式动态捕捉 8

第三章 Intel RealSense技术和Nuitrack 10

3.1Intel RealSense技术概述 10

3.2Intel RealSense产品介绍 10

3.3Intel RealSense D415摄像头 11

3.4Intel RealSense SDK 2.0 12

3.5Nuitrack SDK 12

3.6Nuitrack骨骼系统 13

第四章 深度信息的采集与处理 15

4.1彩色图像和深度图像的对齐 15

4.2深度图像的滤波处理 16

4.3基于深度图像的点云数据 17

第五章 交互系统的设计与实现 18

5.1Unity 3D相关介绍 18

5.2在RGB图像上显示人体骨骼 19

5.3使用骨骼信息驱动人物模型 21

5.3.1间接映射驱动模型 21

5.3.2直接映射驱动模型 22

5.3.3间接映射和直接映射的结果对比 22

5.4用动画表情替换人脸 24

第六章 总结与展望 26

6.1工作总结与实验创新 26

6.2未来展望 26

参考文献 28

致 谢 29

第一章 绪论

1.1三维建模的研究背景及意义

剩余内容已隐藏,请支付后下载全文,论文总字数:26954字

您需要先支付 80元 才能查看全部内容!立即支付

该课题毕业论文、开题报告、外文翻译、程序设计、图纸设计等资料可联系客服协助查找;