基于机器视觉的动态多点手势识别方法

摘要

提出了一种高效的基于HSV颜色空间的多目标检测跟踪方法,实现通过摄像机实时检测跟踪多个指尖目标；定义了一套基于指尖运动轨迹的动态手势模型,并提出了动态手势识别方法；对于两点动态手势,通过BP神经网络进行手势学习和手势识别,而对于模拟鼠标手势和四点动态手势,利用指尖之间相互位置关系进行手势识别.测试结果表明,该方法能够快速、准确的跟踪多个运动的指尖目标并进行动态多点手势识别.%An efficient algorithm for multi-object detecting and tracking based on HSV color space is proposed to detect and track multi fingertips through a camera, a set of dynamic gestures based on the trajectories of fingertips is defined and a method of dynamic gesture recognition is presented: two-point gesture is trained and recognized through BP neural network while four-point gesture is recognized in the light of position relationship between the fingertips. Experimental results show that the proposed method is reliable and efficient for the tracking of the moving fingertips and for the recognition of dynamic gestures.