Radiation imaging from multiple readout of a monolithic scintillator

A monolithic scintillator has advantages over segmented scintillators due to their reduced dead area and straightforward implementation. Previous methods have used a nearest neighbor approach to match multiple photodetector responses to entries in a predetermined look-up table (LUT). We extend this technique by introducing various inverse interpolation techniques, namely kriging, inverse distance weighting, and nonnegative least squares. These interpolation schemes provide greater estimation accuracy at reduced LUT sizes. Various distance metrics are also analyzed which compare simulated photodetector responses with each LUT entry. Slightly higher estimation accuracy is obtained when using the Hellinger f-divergence. We assess the effect of various parameters on the scintillation position estimation accuracy. The estimation error is illustrated by simulating randomly distributed scintillation events for various numbers of detectors, geometries, and numbers of photons. Results indicate that a monolithic scintillator block with multiple non-position sensitive photodetectors can provide sufficient position resolution to be viable for many radiation imaging applications.