Wide-Area Recognition Using Hybrid Motion Stereo

Robots require recognition of the constantly changing environment. Hybrid Motion Stereo combines the two main camera calculation methods, stereo vision and motion stereo, to acquire positional information of the whole field of view. Motion stereo requires the estimation of camera movements to obtain 3D information. Hybrid Motion Stereo applies the publications by stereo vision for estimation of the camera movements. Thus, areas visible in only a single camera are computed using motion stereo, compensating for the disadvantage of stereo vision, which can only compute the areas visible in multiple cameras.