【作者】 吴翔；

【导师】 李文锋；

【作者基本信息】 武汉理工大学 ， 机械设计及理论， 2011， 硕士

【摘要】 随着现代科技的不断发展,不管是在军事领域,还是在日常生活方面,对于虚拟现实技术的要求不断提高。研究基于人体位姿跟踪的虚拟现实技术正在受到广泛的关注,具有重要的研究意义和应用价值。开发设计新型的人体位姿跟踪系统和人体位姿跟踪方法,在游戏娱乐、体育、医疗等领域都具有重大的实用价值。本文对人体位姿跟踪算法进行研究和运用,基于加速度传感器设计了一个用于人体姿态跟踪的虚拟系统,用来演示人体位姿跟踪算法。论文主要包括以下几方面的工作：1.分析了改进的人体位姿跟踪算法的基本原理和使用条件。利用加速度值计算方法采集人体动作信息进行比较,证明利用角度算法对人体位姿的特征记录比利用加速度值更加优越可靠。2.建立了对人体姿态进行跟踪的应用系统模型并改进算法。分析了人体运动的基本规律,利用机械式研究方法的思想设计了人体位姿跟踪系统原型中的模型结构,给出了人体位姿跟踪系统工作的整体框架。3.设计了相应的传感器模块。硬件部分主要包括有传感器信号采集模块、单片机信号处理模块和无线传感器网络无线通信模块。软件部分利用时域的快速信号处理方法。根据需要,将加速度传感器的信号提取出来,在单片机上实现信号处理,并通过无线网络发送到上位机。4.设计开发了人体位姿跟踪系统的人体模型。人体模型是在3Ds MAX软件下建立与优化,模型的控制是以OSG三维渲染引擎为基础,在Visual Studio 2005平台上,采用C++语言实现。5.对开发的人体位姿跟踪原型系统进行实验与分析评估。设计了一组比较实验,分别利用加速度和角度来记录人体运动的位姿信息,并进行比较,分析得出利用角度信息记录比较准确。对人体的四肢运动分别进行可视化跟踪实验,分析了利用角度算法来进行实体跟踪遇到的一些问题。更多还原

【Abstract】 With the development of modern science and technology, application requirements for virtual reality technology are rapidly increasing in both military and our daily life. Research on the virtual reality technology based on body posture tracking receives widespread attention, which is an important research with significance and application value. The development and design of new body posture tracking system and method has great practical value in the game entertainment, sports, medical and other fields.This thesis focuses on the body posture tracking algorithm and its application. A body posture tracking system used accelerometer is developed to demonstrate the body posture tracking algorithm. Main content and innovations are as follows:1. Analyze body posture tracking algorithms. Its basic principles and conditions are introduced. It is proved that new algorithm using the characteristics of the human body posture recorded is excellent and more reliable than the use of accelerometer.2. Build the application model of body posture tracking and improve its algorithm. Through analyzing the basic laws of human motion, the mechanical design methods is introduced to build the model structure of body positure tracking system prototype, and then give out the entire system framework.3. Design sensor module. Hardware components includes sensor signal acquisition module, signal processing module and transceiver module. Software components realize fast signal processing by using time-domain methods. To correspond with hardware, accelerometer signal will be sampled by the microcontroller. After processing, it is transmitted to a host computer.4. Build body posture model for body posture tracking system. This model is established under the 3Ds MAX software. The model optimization, which is based on three-dimensional rendering engine based on OSG, is realized by C++ language on the platform of Visual Studio 2005.5. Experiments and evaluation for the body posture tracking system prototype to prove the possibility of body tracking using algorithm based on angles. A special experiment is designed to compare the algorithm based on angles with algorithm based on acceleration. The result reveals that algorithm based on angles has better accuracy. Another experiment is to record the body movement for visual tracking, and analyzes problems of using algorithm based on angles for body tracking.更多还原