由光流确定刚体的三维运动参数和表面形状

Determination of the 3-D Motion Parameters and Surface Shape of a Rigid Body from Optical Flow

本文提出了一种由光流估计刚体的三维运动参数的方法。该方法具有以下几个特点;1.该方法是线性的,不需迭代求解;2.建立的运动估计方程所用的未知数个数降至五个——刚好是可恢复的运动参数个数,大大缩短运动估计所用的时间;3.引入最小二乘法使算法对噪声具有平滑作用。最后讨论由运动参数恢复可见表面形状的可能性。

In this paper, a method is presented for estimating 3-D motion parameters of a rigid body. This method has following features: 1. It is a linear method in which the interactive solving is not needed. 2. The equations for motion estimation contain only five unknowns-the number of the recoverable motion parameters. The time cost of motion estimation is largely reduced. 3. The Least-Square method is used so that the algorithm can be applied to the image sequence with noise, recovering the shape of visible surface from motion parameters is discussed.