[Back]

E. Gröller, A. Amirkhanov et al.:
"Fuzzy CT Metrology: Dimensional Measurements on Uncertain Data";
Talk: Spring Conference on Computer Graphics, Smolenice, Slovakia; 2013-05-01 - 2013-05-03; in: "29th Spring Conference on Computer Graphics 2013", (2013), ISBN: 978-80-223-3377-1; 93 - 101.

Metrology through geometric dimensioning and tolerancing is an important instrument applied for industrial manufacturing and quality control. Typically tactile or optical coordinate measurement machines (CMMs) are used to perform dimensional measurements. In recent years industrial 3D X-ray computed tomography (3DXCT) has been increasingly applied for metrology due to the development of XCT systems with higher accuracy and their ability to capture both internal and external structures of a specimen within one scan. Using 3DXCT the location of the specimen surface is estimated based on the scanned attenuation coefficients. As opposed to tactile or optical measurement techniques, the surface is not explicit and implies a certain positional uncertainty depending on artifacts and noise in the scan data and the used surface extraction algorithm. Moreover, conventional XCT measurement software does not consider uncertainty in the data. In this work we present techniques which account for uncertainty arising in the XCT metrology data flow. Our technique provides the domain experts with uncertainty visualizations, which extend the XCT metrology workflow on different levels. The developed techniques are integrated into a tool utilizing linked views, smart 3D tolerance tagging and plotting functionalities. The presented system is capable of visualizing the uncertainty of measurements on various levels-of-detail. Commonly known geometric tolerance indications are provided as smart tolerance tags. Finally, we incorporate the uncertainty of the data as a context in commonly used measurement plots. The proposed techniques provide an augmented insight into the reliability of geometric tolerances while maintaining the daily workflow of domain specialists, giving the user additional information on the nature of areas with high uncertainty. The presented techniques are evaluated based on domain experts feedback in collaboration with our company partners.

metrology, uncertainty visualization, level-of-details, industrial 3D computed tomography

http://publik.tuwien.ac.at/files/PubDat_223587.pdf