计算机视觉研究中,为了实施对人运动图像的跟踪,本文提出了基于变化区域的对象刚体标注法和语义模型跟踪法,通过对标注的刚体特征点在空间位移的估算,结合特征点在水平面上投影所形成的特征码来构成刚体的运动语义数学模型,从语义模型来计算特征点的运动位置,从而实现对人运动的跟踪,对于人的复杂运动描述也具有参考价值。
This paper discussed about the human tracking of computer vision. From detected the changed region in the image, the object of changed region was labeled as rigid body which selected from model database. The eigendots had been set on it. During the motion, the displacement of the moving eigendot in space had been computed. Added by the moving eigendot pro- jected to the water level and constructed the eigenvalue code, the mathematic model has been built on displacement and eigenvalue. The computing position of eigendot was being used to present the human motion tracking. It was also benefits to complex motion description.