Tomographic reconstruction for the IAEA PGET detector: a competition entry

Abstract

The international atomic energy agency (IAEA) held a challenge about tomographic reconstruction and analysis algorithms for its PGET detector. This device measures activity levels in burnt nuclear fuel pins. The algorithm is asked to identify misplaced rods too. The challenge is complicated due to the high cross section for gamma photons in the energy range of interest of uranium, attenuating signal from rods at the center of the assembly. Compton scattering will also remove spatial information from most detected events. This document describes my solution to the challenge. To reconstruct images, algorithm from the nuclear medicine field have been used. Their decade long use proved their robustness and lack of biased. Subsequent improvements are used for attenuation correction. This is performed using an intermediate reconstruction where missing rods are identified. This knowledge is exploited to generate the last attenuation map, which provides the correction for the final quantitative reconstruction. Compton scatter correction is performed by subtracting counts in different energy windows. Performances were tested on training data from 60Co mock assemblies. On such data 100% accuracy for missing rod determination was achieved.