Registration Of A Moving Rigid Object Using A Stereoscopic Vision Setup

This paper addresses the problem of tridimensional registration of a moving rigid object. Matching, tracking and 3D reconstruction of feature points by a stereoscopic vision setup allows the computation of the homogeneous transformation matrix linking two consecutive scene captures. Robustness to errors is provided by the scene rigidity constraint. Accumulation of error is compensated through loop detection in the calculated camera positions.