Voxel dataset noise reduction algorithm based on CMM reference measurements for geometrical tolerance analysis of prismatic test pieces

Computed tomography for dimensional metrology has been introduced in the quality control loop for about a decade. To make the tolerance verification consistent, GD&T evaluation programs have been developed, which make use of CT point clouds generated through segmentation. This important step in a CT measurement sequence is investigated under the perspectives of tolerance analysis and measurement uncertainty. Based on empirical master part reference measurements, voxel dataset noise can be rationally filtered out by applying point-based reduction algorithms. Their significance on dimensional and geometrical characteristics’ measurements is experimentally tested and discussed in this work.