A two-stage method for selective modeling of unknown teleoperation environments

A semi-automatic method for the construction of geometric object models in unknown teleoperation environments is proposed. The problem is divided in two parts: detection and designation of the target and target model generation, In the first, interactive phase of the method the operator observes and designates some feature points on at least two different perspective images of the target environment. In the second phase range measurements are automatically generated across the estimated target 3D-sub-volume. Finally, a surface model is reconstructed from the measurement data.<
>