Time series evaluation of radiation portal monitor data for point source discrimination

Time series of data from radiation portal monitors are evaluated for radioactive sources by comparing background to vehicle spectra over time with a “spectral-distance” metric, isolating the contribution of anomalous radiation. This may diminish systematic fluctuations from vehicular background attenuation and allow time-shape filtering for discriminating compact sources. To examine this, a wavelet function similar in size to the expected source profile filters the spectral-distance output. Spectra from chosen isotopes are injected into data from a U.S. port of entry. The resulting data are analyzed with gross-counting, spectral-distance, and spatial algorithms. Combined spectral/spatial filtering is shown to enhance sensitivity and discrimination of compact versus distributed sources.