基于微软Kinect体感游戏控制器的人体识别方法研究
关键字：Dijkstra算法，稀疏Hausdorff距离
Title Research on the Real-time Full Body Pose
Reconstruction based on Microsoft Kinect Sensor
Abstract
In recent years, depth cameras have become a widely available sensor type
that captures depth images at real- time frame rates. Even though recent
approaches have shown that 3D pose estimation from monocular 2.5D depth images
has become feasible, there are still challenging problems. In this paper, we
present an efficient and robust pose estimation framework for tracking
full-body motions from a single depth image stream. Following a data-driven
hybrid strategy that combines local optimization with global retrieval
techniques, we contribute several technical improvements compared to previous
approaches. In particular, we introduce a variant of Dijkstra’s algorithm
to efficiently extract pose features from the depth data and describe a fusion
scheme based on an efficiently computable sparse Hausdorff distance to combine
local and global pose estimates. Our experiments show that the combination
of these techniques facilitates real-time tracking with stable results，making
it applicable to a wide range of interactive scenarios.
Keywords : Dijkstra’s algorithm，Hausdorff distance.
目录
1 绪论 6
1．1 课题背景 . 6
1．2 难点问题分析 6
1．3 本文的研究内容与章节安排 . 6
2 Kinect 原理介绍 7
2．1 引言 7
2．2 Kinect 技术介绍 . 7
2．3 Kinect 体感控制器的功能 . 8
2.3.1 使用红外定位 8
2.3.2 多项额外功能 9
2.3.3 配备自有界面 9
2．4 Kinect 体感控制器各功能部件介绍 . 9
2.4.1 红外线发射器和红外线/VGA 摄像头组 9
2.4.2 麦克风10
2.4.3 各种芯片及电路板.10
2.4.4 零件汇总 12
3 Kinect 自然用户界面技术 NUI12
3.1 Kinect Windows 构架 12
3.2 NUI API .14
3.3 NUI API 的初始化 .15
3.3.1 传感器的枚举与访问 15
3.3.2 初始化选项 17
3.4 NUI 图像流概要 .17
3.4.1 色彩数据 18
3.4.2 深度数据 18
3.4.3 玩家段数据 19
3.4.4 检索图像信息 .20
3.5 NUI 骨骼追踪.25
3.5.1 检索骨骼信息 .25
3.5.2 主动及被动追踪 .26
3.5.3 NUI 转换 27
4 实验总体框架 .29
4.1 NUI API 封装 .30
4.1.1 初始化 .30
4.1.2 关闭 Kinect 传感器 31
4.1.3 对图像数据流的采集 32
4.2 人体模型 .33
4.3 特征提取 .35
4.4 体态数据库 35
4.4.1 数据库构建 35
4.4.2 数据库查找 36