Adaptive sampling plan design methodology for reverse engineering acquisition

Abstract

Reverse engineering is a technology which generates a virtual representation of an existing part based on point data acquired through measuring techniques. Different technologies can be employed to obtain a virtual representation of a physical model, but the use of a solution (3D scanner) rather than another provides significantly different results since the available 3D scanners are characterised by different performances (resolution, accuracy, ...). However, even if great attention were focussed on the selection of the most appropriate 3D scanner device, this would not be enough to assure the achievement of a consistent virtual representation of the physical model. The selection of the most suitable 3D scanner can contribute to rendering the point acquisition more accurate, but it is not able to assure an efficient point distribution in terms of numbers and locations. These two parameters are part of the acquisition strategy, which can be implemented only after having decided which the 3D scanner is to be used. In order to support the next steps of the reverse engineering cycle (segmentation, fitting, ...), the acquisition phase should provide an organised point cloud, which can be obtained through a consistent sampling plan. For this reason, this paper aims at proposing a methodology for defining a selective sampling plan, whose grid dimensions are related to the complexity of the analysed local surface regions.