THIS PAPER PRESENTS A REAL-TIME ALGORITHM FOR IDENTIFYING AND LOCATING PARTS FOR ROBOT PICK-UP. THE ALGORITHM IS BASED ON THE SEARCH FOR BEST FEATURE POINTS AND HAS BEEN DEVELOPED FOR IDENTIFYING AND LOCATING INDUSTRIAL PARTS. BOTH THE FEATURE POINTS IN THE DATABASE AND THE OBJECT IN THE SCENE CAN BE CHARACTERIZED USING THE SAME ALGORITHM. BY EFFECTIVELY EXCLUDING NOISY FEATURE POINTS FROM THE COMPUTATION OF THE PART'S LOCATION AND ORIENTATION, THE ALGORITHM CAN DEAL WITH PARTS WITH PARTIAL OVERLAPPING, MISSING OR UNWANTED FEATURES, AND SCALE CHANGES. THE COMPUTATION LOAD CAN BE MINIMIZED IN AN OFF-LINE PROCESS FOR REAL-TIME IMPLEMENTATION USING A LOW-COST MACHINE, SUCH AS A PERSONAL COMPUTER.
| Number of Pages : | 17 |
| Published : | 06/01/1992 |
