虚拟现实训练系统中基于手势的人机交互  

Hand Gesture-based Interaction in Virtual Reality Training System

作　　者：徐德友[1] 

机构地区：[1]炮兵学院南京分院新弹药研究中心,江苏南京211132

出　　处：《系统仿真学报》2006年第z2期386-389,共4页Journal of System Simulation

基　　金：江苏省自然科学基金资助项目(BK2006004)

摘　　要：给出了虚拟手建模和数据手套校正的方法和步骤。通过数据手套与位置跟踪设备的结合使用,得到现实世界中手的姿势以及位置,并以此驱动虚拟世界中的虚拟手进行交互操作。运用神经网络和逻辑组合的方法分别对定义的静态手势和动态命令进行识别,形成相应的操作命令。为验证所提方法的有效性和实用性,将基于手势的人机交互技术应用于某型自行火炮的虚拟驾驶训练中,受训者可以像在真实世界中一样操纵虚拟训练环境中的虚拟物体和部件,取得了满意的训练效果。The methods of virtual hand modeling and dataglove calibration were proposed firstly. The gestures and positions of the hand in real world, which were tracked with the datatglove and motion tracking device, were used to drive the virtual hand in virtual environment. Neural network and logic combination was used to recognize the static hand gestures and dynamic commands. The required computer manipulation commands were got according to the recognition results. The efficiency and practicability of the presented methods were validated by applying it to the virtual driving training system of self-propelled gun (SPG). In the system, the users can manipulate the virtual objects and components as if they are in real world.

关 键 词：虚拟现实 人机交互 手势识别 神经网络 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]