摘要随着科学技术和人类认识世界需求的不断提高，传统的机器视觉已经不能满足人们对于物体识别的要求。与灰度图像相比，深度图像具有物体三维特征。由于深度图像不受光线照射方向及物体表面的发散特性的影响，而且不存在阴影，所以可以更准确地表现物体目标的三维深度信息。64195

本论文研究的内容包括使用OpenGL进行简单的三维模拟环境搭建，并以此为实验环境，进行简单的深度数据的分析研究，使用OpenNI对现实环境进行深度数据采集，并做去噪平滑等处理以便研究分析。本课题成功实现了虚拟环境和现实环境的用于室内导航的走廊识别，并实现了基于虚拟环境的三维重建。

毕业论文关键词  深度图像分析  三维模拟  OpenNI   3D重建 

毕业设计（论文）外文摘要

Title   Corridors Identified Based on Kinect Depth Data  and Simulation of OpenGL 3D  Simulated Environment                 

Abstract

As science and technology and human understanding of the world demand continues to increase, the traditional machine vision has been unable to meet the requirements for object recognition. Compared with the gray-scale image, depth image has three-dimensional object features. there is no shadow and the influence of the direction of the light in the image and the surface characteristics of the pergent is ignored in the depth image, so it can provide a more accurate representation of the three-dimensional depth information of the target object.

The contents of this research include ,building a simple OpenGL 3D  environment , and use it as a laboratory environment, a simple depth data analysis.Using OpenNI to get frames of depth data from reality environment and processing in order to do denoising smoothing for later analysis. This project implement corridors identification of indoor navigation both in mock and real environment,also do some 3D reconstruction using a mock environment.

Keywords  analysis of depth image   three-dimensional simulation   OpenNI 3D image reconstruction

1. 绪论 1

1.1 深度数据分析的概况 1

1.2 机器人室内导航 2

1.3 研究内容 3

2 配置MFC环境下的OpenGL 4

2.1 OpenGL简介 4

2.2 MFC环境下的OpenGL配置 5

3. 三维模拟环境的搭建与深度数据的获取 8

3.1 三维模拟环境初始化参数配置 8

3.2构建三维模拟环境 10

3.3 模拟环境深度数据的获取 12

3.4 摄像机漫游 13

4. 基于深度曲线的走廊识别 16

4.1走廊识别 16

4.2 走廊识别的算法缺陷 20

5. 以Kinect基于OpenNI分析现实曲线 22

5.1 OpenNI简介 22

5.2 现实环境下深度曲线 24

5.3 走廊识别算法的改进 27

6深度图像在3D重建中应用的讨论 30

6.1  3D重建简介 30

6.2 简单的3D重建及原理