适用于机器人视觉的手势识别系统

手势是一种高效的人机交互和设备控制的方式,基于视觉的手势识别是人机交互、模式识别等领域的一个富有挑战性的研究课题。文章提出并实现了一个可用于与机器人交互的静态手势检测和识别系统。该系统用摇动检测的方法定位人手;用基于现场采样得到的肤色模型进行手的分割;用简化并改进的CAMSHIFT算法对手势进行跟踪;最后用模式识别的方法提取简单特征进行识别。实验证明,该系统快速、稳定而有效。

A Gesture Recognition System Used on Robot Vision

Gesture is a kind of effective human-computer interaction and device control techniques.Vision-based recognition of hand gestures is a challenging project in computer science.This paper puts forward and realized a practical gesture system which can be used in controlling robot and HCI.It locates hand by waving detection,segment gesture using present skin color model,track gesture using simplified CAMSHIFT tracker and recognize gesture by simple features finally.Experiments show that the system proposed is fast,stable and effective.