摘要
随着虚拟现实技术的不断发展和成熟,人们已经不满足于简单的鼠标键盘交互。利用人手与虚拟环境进行交互,可以提高计算机的可操作性。但是对手的跟踪比较复杂,因为手是一个高自由度的关节结构。在分析手的运动学模型以及总结手运动约束关系的基础上,利用双目立体视觉系统,即用两台摄像头同时采集手的图像,通过识别在手上设置的关键点位置,结合设计的简化计算模型,推算出手指的运动位置,从而在虚拟环境中再现手的姿态。该方法降低了对硬件的要求,没有数据手套运动不灵活、沉重等缺点,计算模型的简化能够提高识别的速度,在与虚拟现实环境的交互中得到了较好的应用。
As the continuous development and maturity of VR technology, people are not satisfied with simple interaction by using mouse and keyboard. Using human hand to interact with computer could increase the maneuverability of computer. But the tracking of hand is complex, because hand is a high degree of freedom articulate mechanism. Binocular solid vision system, which is making use of two cameras to catch the movement of hand and combines with the simplified compute model to conclude the fingers' location after recognizing the characteristic points on hand, is used based on the analysis of hand kinematics model and the hand constraints. In the end, the gesture of Visual Hand is rendered. This method reduces the demand for hardware. There is no inconvenience that the disadvantage of Data Glove brings about. Also the simplified compute model could increase the speed of recognition, so it has good application in the interaction with VR environment.
出处
《工程图学学报》
CSCD
北大核心
2007年第4期30-35,共6页
Journal of Engineering Graphics
基金
国家自然科学基金资助项目(60573174)
合肥工业大学学生创新基金资助项目(XS05003)
关键词
计算机应用
双目立体视觉
运动约束
虚拟手
动画
computer application
binocular solid vision
movement constraints
visual hand
animating