Numeric Prediction of the Detail Visibility in Industrial X-Ray Computed Tomography by Human Observers

Abstract

Industrial Computed Tomography (iCT) is applied in industry for flaw detection, flaw evaluation and dimensional measurement. This requires the correct experimental system settings for sufficient visibility and detectability of details and structure elements. This is an essential tool for long term monitoring of the detail sensitivity of CT scanners. The visibility of indications by human observers on a monitor in cross sectional 2D CT-images can be estimated from the square root of the visible flaw area, the Contrast to Noise Ratio (CNR) and the Modulation Transfer Function (MTF). The ASTM guide E 1441 describes a more detailed procedure for the determination of the minimum contrast for the visibility of flaws based on three essential functions for the prediction of the visibility of small circular indications in iCT slice images. This is the Contrast Discrimination Function (CDF), the Modulation Transfer Function (MTF) as described in the latest revision of ASTM E 1695, and the Contrast Detail Diagram (CDD). All these functions analyse the contrast as function of the spatial frequency in digital 2D slice images. The functions do not describe, how to calculate the visibility limit for human observers. The concept for the automated calculation of the visibility limit of circular indications in reconstructed slice images is discussed in this paper. It is finally determined from the intersection point of the MTF with the Contrast Detail Diagram, which is the combination of CDF and MTF, and a physiological factor c. The new measurement procedures for the prediction of the detail visibility by MTF and CDD was tested with test phantoms and be verified by modelling and measurements. A Round Robin test was conducted with more than 10 parties to verify the visibility formula and the procedure for determination of the visibility limit from the combination of these functions. A form factor is considered to compare cylinder holes with pore indications. Conclusions will be reported and recommendations will be given for the determination of the correct physiological factor c.