Parallel optimization-based spectral transformation for detection and classification of buried radioactive materials

In this paper, we investigate detection and classification of buried radioactive materials of interest using data collected by a Sodium Iodide (NaI) detector with short sensor dwell time (i.e., less than or equal to 1 s). Due to the sparseness and randomness of a gamma spectrum, four different spectral transforms are implemented as mechanisms for background normalization and feature extraction. These spectral transforms are based on binned energy windows determined by a particle swarm optimization (PSO) method. However, the overall process is time-consuming. Since parallel computation is an appropriate approach to reduce the computation burden of the PSO-based search process, the parallel implementation on cluster for optimal bin partition is proposed in this paper. The speedup performance and resulting detection and classification performance are investigated.