ABSTRACT
The problem of automated fettling was discussed by Balendran(1992). From his work a project has evolved looking at the problem of 3D object recognition and free-form surface matching. The aim of this work is to develop a solution for use in a manufacturing environment, specifically, but not limited to, automated fettling. The initial work has centred on the development of the front-end. The following requirements for the front-end were determined:-
• Real-time data collection rates, • The capability to measure a wide range of object complexities, • Must not be reliant on the nature of the objects surface, • View point independent data collection, • Small enough to be mounted on a robot arm, • Robust.
