D. Cobzaş, M. Jägersand, and H. Zhang
Computer vision, image-based modelling, robot navigation, range-intensity data registration, panoramic model, predictive display
Robot tele-operation is significantly degraded by delays in the operator visual feedback. The authors present a cylindrical image- depth model that can be easily acquired at the remote work site and then used to support both robot localization and predictive display. They present an implementation for a mobile robot system where synthesized immediate visual feedback replaces the delayed real images. Experimentally, they found that a predicted view could be rendered with a geometric viewpoint error of no more than 2.4 cm/1.2 degrees in a 100 m2 work scene.
Important Links:
Go Back
