Multi-sensor data fusion for realistic and accurate 3d reconstruction

In this paper, the problem of multi-sensor data fusion for 3D dimensional reconstruction/assessment of manufactured parts is investigated. For this, a method relying on cooperation of two acquisition/processing techniques (stereoscopic and structured light sensors) is presented. The stereoscopic system reconstructs the outline of the object, whereas the structured light system represents the workpiece as points that characterize its 3D surfaces. The developed approach exploits the complementarity of the two acquired data sets. To represent the whole set of measurements in a common reference frame, a fusion method is developed based on the geometric characteristics of the workpiece used to define a cost function, which allows perfect alignment of the two point sets. This is further favored by use of a camera common to the two sensory systems. Experimental results demonstrate that the proposed method outputs results, which can be meaningfully used in applications such as quality control.