Traceable radiographic scale calibration of dimensional CT

Speaker:
Illemann, Jens; Physikalisch-Technische Bundesanstalt; Germany

Authors:
Illemann, J.; Physikalisch-Technische Bundesanstalt; Germany
Schulz, D.; Physikalisch-Technische Bundesanstalt; Germany
Neuschaefer-Rube, U.; Physikalisch-Technische Bundesanstalt; Germany

ID: ECNDT-0175-2018
Download: PDF
Session: CT-Methods 1
Room: G3
Date: 2018-06-12
Time: 10:50 - 11:10

Computed tomography (CT) is based on digital radiography. In the last 15 years the use of CT has been established as a coordinate measuring system for dimensional measurements. In order to reach a small uncertainty, correct scaling of the underlying radiography is required. Due to the finite volume of the X ray source, source degradation, the finite thickness of the detector layer and additional mechanical deviations, a typical industrial cone-beam CT has relative scale deviations of 10-4 or more. Former work by the authors shows that, by using a thin grid-like structured metal layer, the magnification can be easily determined with a scale resolution better than 10-6. For this purpose, two radiographs are imaged in 180 degrees opposite directions on the rotary stage of the CT system. The harmonic average of the two magnifications is the magnification belonging to the rotary axis’ centre.
In this article, investigations on the development of a radiographic scale standard are presented. Thin metal foils with holes at equidistant grid positions were fabricated by means of different methods and calibrated on an optical coordinate measuring machine. Depending on the production process, specific contours arise that are imaged differently by the radiographical and the optical method. Based on statistical evaluations of the congruence of both results, the hole centres are determined to better than 1 µm. Further simulations allow an uncertainty to be specified for the radiographic scale determination that is traced back to optical calibration. The result is an SI-traceable determination of the radiographic magnification with a relative uncertainty of a few 10-6.