Low Count Anomaly Detection at Large Standoff Distances

• Published in: IEEE transactions on nuclear science Vol. 57; no. 1; pp. 309 - 316
• Main Authors: Pfund, D.M., Jarman, K.D., Milbrath, B.D., Kiff, S.D., Sidor, D.E.
• Format: Journal Article
• Language: English
• Published: New York IEEE 01-02-2010; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Anomalies; Anomaly detection; Background radiation; Biomedical imaging; Counting; Cutoff frequency; Detection algorithms; Detectors; Filters; Gamma rays; gamma-ray spectroscopy; Imaging; Isotopes; radiation monitoring; Radioactive materials; Searching; Spectrometers; Spectroscopy; Working environment noise
• ISSN: 0018-9499; 1558-1578
• DOI: 10.1109/TNS.2009.2035805

Searching for hidden illicit sources of gamma radiation in an urban environment is difficult. Background radiation profiles are variable and cluttered with transient acquisitions from naturally occurring radioactive materials and medical isotopes. Potentially threatening sources likely will be nearly hidden in this noise and encountered at high standoff distances and low threat count rates. We discuss an anomaly detection algorithm that characterizes low count sources as threatening or non-threatening and operates well in the presence of high benign source variability. We discuss the algorithm parameters needed to reliably find sources both close to the detector and far away from it. These parameters include the cutoff frequencies of background tracking filters and the integration time of the spectrometer. This work is part of the development of the Standoff Radiation Imaging System (SORIS) as part of DNDO's Standoff Radiation Detection System Advanced Technology Demonstration (SORDS-ATD) program.