| Home | Projects | People | Publications | Links |
|
Three-Dimensional Robot Vision N. Ahuja, N. Bridwell, C. Nash and T. S. Huang Describes an ongoing research project in the extraction, representation and utilization of 3-D information obtained from spatial or temporal stereo image input. As the first step, information about occupancy or unoccupancy of space is derived from stereo images. A volumetric (octree) representation is then derived for the occupied space. The known model of the objects in a scene is used to predict collisions among moving objects, and to plan safe and efficient trajectories. Full Text: NA |