Abstract
This paper presents a vision system that finds and measures the location of 3D structures with respect to a CAD-model. The integration of a CAD-model to visual measurement and direct feedback of measurement results to the CAD is a key aspect. For the extraction of basic visual cues, independent and complementary modules are envisaged. The goal is that of navigating a legged robot into a ship structure using the pose estimated from visual landmarks extracted from the CAD-model. These are tracked in real-time by the vision system and are matched to the CAD-model. For the implementation of the vision system, commercial off-the-shelf parts were used along with a custom designed robot stereo head. The communication with other modules is done using simple ASCII commands on a fiber channel network with standard TCP/IP protocol. This allows easy debugging, straightforward development of applications software with little effort, and very high data rates.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
