An efficient method for fully automatic 3D digitization of unknown objects
Computers in Industry
| Hi-index | 0.01 |
The paper addresses an implementation issue of a robotized reverse engineering platform for not modelled objects. The RE platform is composed of three major components: a 6 d.o.f. industrial robot, a dual-camera laser scanner and a digitally controlled turntable on which the object is placed. One important objective of this work is to develop suitable scanning strategies and robot motion patterns for automatic sensor guiding and acquisition of 3D surface data of objects to be modelled. The proposed strategy implies the synchronization of the robot movements with the rotation of the turntable, in order to avoid collisions or axis out of range and to keep, if possible, the robot in the central area of its workspace. For each unique picture taking point, the adequate turntable angle and robot configuration will be computed. A 6 d.o.f vertical articulated robot simulator was developed to generate the set of necessary robot configurations and turntable angles corresponding to the scanning strategies. The simulator is used to test whether a scanning path is feasible from the kinematics point of view. The scanning strategy consists of two stages: first, predefined scanning paths are followed, based on offline planned motion patterns; secondly, evaluation of the acquired 3D data may eventually lead to new scanning paths with different orientations of the hand-held laser scanner to obtain complete 3D information of the hidden details of the object’s surface.