摘要旗语是目前海上通信不可缺少的一种通信方式。研发出自动识别旗语的方法，使计算机能够理解旗语具有重要的实用意义。本文主要描述了一个基于Kinect的旗语识别系统的开发过程。首先介绍了旗语、虚拟学习环境及Kinect的研究背景，之后较为详细地介绍了该系统所使用的方法和工具，最后描述了具体的实现步骤。18997
本系统使用Kinect结合Unity 3D实现，以Kinect捕捉的用户的骨骼数据控制虚拟场景中的人物。用户可以通过模仿场景中模特的动作学习旗语，也可以测试自己对旗语的掌握程度。初步实验表明，该系统可满足模拟训练的需求，作为旗语学习系统可能对初学者学习旗语有帮助。但也发现本系统存在很大的改善空间。
关键词 Kinect 手势识别 旗语 游戏 虚拟学习环境
毕业设计说明书（毕业论文）外文摘要
Title The System Of Gesture Recognition Based On Kinect
Abstract
Flag semaphore is indispensable to maritime communication nowadays. It has important practical significance to develop automatic semaphore identification method and make the computer to understand the semaphores. This paper describes the development process of a Kinect-based semaphore recognition system. First introduced the background of semaphore, virtual learning environment and Kinect, after which a more detailed description of the methods and tools used in the system, and finally describe the specific implementation steps.
The system was achieved through the Kinect and Unity 3D, it use the Skeleton Data of Kinect to control the character in the virtual scene. Users can learn semaphores by following model’s actions, they can also test their mastery of semaphores. Preliminary experiments indicate that the system can meet the needs of simulation training, and may be helpful for beginners to learn semaphore as a semaphore learning system. But it was also found that the system has a lot of room for improvement.
Keywords Kinect; gesture recognition; flag semaphore; game; virtual learning environment
目 录
1 绪论 1
1.1 相关背景 1
1.1.1 旗语 1
1.1.2 虚拟学习环境 1
1.2 研究目的 2
2 采用的工具和方法 3
2.1 Kinect原理简介 3
2.1.1 Kinect的主要硬件 3
2.1.2 获得深度数据 4
2.1.3 生成骨骼图 4
2.1.4 识别人体关节 4
2.1 游戏引擎Unity 3D 5
2.2.1 游戏引擎 5
2.2.2 Unity的动画系统 6
2.2 MikuMikuDance 7
2.3 日本旗语体系 8
3 系统构建 10
3.1 环境搭建 10
3.2 插件的使用 10
3.2.1 Kinect Wrapper 10
3.2.2 MMD4Mecanim 11
3.3 手势识别 14
3.3.1 单手的姿势 15
3.3.2 识别静态手势 17
3.4 增加教学演示 18
3.5 系统完善 19
4 问题与反思 20
结 论 22
致 谢 23
参考文献24
1 绪论
1.1 相关背景
1.1.1 旗语 基于Kinect的旗语识别系统设计:http://www.751com.cn/jisuanji/lunwen_40330.html