Multiple-source detection and identification using 3D-position-sensitive semiconductor detectors

In some applications, one wishes to simultaneously detect, identify, and localize multiple point sources of the same or different isotope in an unknown or known background environment. In this work, we use the Multifamily Likelihood Ratio Test (MFLRT) to detect the presence of a known number of point sources and, optionally, to estimate the number of sources present, using 4π Compton imaging with 3D-position-sensitive CdZnTe detectors. In addition, the sources´ intensities, positions, and isotopic identities (from a library) are estimated. No prior knowledge of the background spectral distribution is necessary, and point-source isotopes with spectral lines also present in the background can be detected. If available, known background spectral shape and intensity can be included. Spatially, the background is assumed to be smooth with variation of less than a factor of two. Experimental results are shown for an 18-detector CdZnTe array system as a function of false-alarm rates.