Direct robot motion vision by fixation

摘要

A direct method, called fixation, for solving the general motion vision problem is introduced. This method results in a constraint equation between translational and rotational velocities that in combination with the brightness-change constraint equation solves the general motion vision problem, arbitrary motion relative to an arbitrary rigid scene. Avoiding correspondence and optical flow has been the motivation behind the direct methods. Recently, direct motion vision methods have used the brightness-change constraint equation for solving special cases such as known depth, pure translation or known rotation, pure rotation and planar world. In contrast to those solutions, the fixation method does not put such severe restrictions on the motion or the environment. Partial implementation of the fixation method shows encouraging results.