Contaminant detection in non-destructive testing using a CZT photon-counting detector

Abstract

With recent advances in the growth of CdZnTe (CZT) sensors, high-flux photon-counting detectors (PCDs) have begun to see more use commercially in non-destructive testing (NDT). One such application is food inspection, where radiography is currently used to detect undesirable contaminants introduced in the production and packaging processes. PCDs can offer better detection than conventional radiography due to the preservation of energy data by analyzing the pulse height of each x-ray detection and sorting the x-ray into one of a number of energy bins. However, there are a number of parameters that must be explored in order to offer efficient and efficacious detection of contaminants. Here, two such parameters were investigated in a phantom study with an 8×24 mm2 CZT detector for a number of common contaminant materials. The detectability of contaminants was evaluated based on their contrast-to-noise ratio (CNR) in 2D transmission images. First, the energy bin demonstrating the highest CNR for each contaminant material was found by adjusting the threshold energies defining the edges of the bin. Second, various pixel binning schemes were utilized to lower noise and investigate the effect on the detectability based on the size of contaminants. CNR was maximized for pixel binning that corresponded to the approximate size of the contaminant objects in x-ray images.