Neural Network Approaches for Mobile Spectroscopic Gamma-Ray Source Detection

Abstract

Artificial neural networks (ANNs) for performing spectroscopic gamma-ray source identification have been previously introduced, primarily for applications in controlled laboratory settings. To understand the utility of these methods in scenarios and environments more relevant to nuclear safety and security, this work examines the use of ANNs for mobile detection, which involves highly variable gamma-ray background, low signal-to-noise ratio measurements, and low false alarm rates. Simulated data from a 2” × 4” × 16” NaI(Tl) detector are used in this work for demonstrating these concepts, and the minimum detectable activity (MDA) is used as a performance metric in assessing model performance.In addition to examining simultaneous detection and identification, binary spectral anomaly detection using autoencoders is introduced in this work, and benchmarked using detection methods based on Non-negative Matrix Factorization (NMF) and Principal Component Analysis (PCA). On average, the autoencoder provides a 12% and 23% improvement over NMF- and PCA-based detection methods, respectively. Additionally, source identification using ANNs is extended to leverage temporal dynamics by means of recurrent neural networks, and these time-dependent models outperform their time-independent counterparts by 17% for the analysis examined here. The paper concludes with a discussion on tradeoffs between the ANN-based approaches and the benchmark methods examined here.

Implementation of Novel Evolutional Algorithm for 3-Dimensional Radiation Mapping and Gamma-Field Reconstruction within the Chornobyl Sarcophagus

“…Although technologies have seen continued advancement over recent decades, following the March 2011 accident at Japan's Fukushima Daiichi Nuclear Power Plant (FDNPP)notable advancements in the unmanned aerial vehicle (UAV) [9][10][11], unmanned ground vehicle (UGV) [12,13], and static/mobile distributed detection systems [14] were realized. Even years after this driver, progress continues across radiation detection, localization, and mapping, whether in underpinning detector materials research (e.g., novel plastics, high-dose semiconductors, dual gammaneutron scintillators) [15][16][17][18]; innovative, autonomous, and miniaturized deployment mechanisms [19][20][21]; or sensor-fusion/data visualization methodologies, progressing from 2D to 3D scenarios [22][23][24][25][26][27]. This research is not just occurring at a small number of institutions and laboratories, but it is a promising and increasingly cross-disciplinary area of active research around the world, applicable for nuclear plant decommissioning, nuclear security, and safeguards applications, whether or not in accident response [25].…”

An Ensemble Approach to Computationally Efficient Radiological Anomaly Detection and Isotope Identification

“…Table I shows the list of source types and activities used for the dataset. The same dataset was also used for benchmarking by Bilton et al, and more detailed data generation procedure is described in [14]. The algorithms were assessed on the augmented dataset in terms of receiver operating characteristic (ROC) curves and minimum detectable amount (MDA).…”

“…Others make use of specific energy windows within which the source and background count contributions are estimated using the spectrum outside the windows [13]. Additionally, some recent works utilize modern machine learning techniques, such as deep neural networks (DNN) [14], [15] and Gaussian processes (GP) [16].…”

Explaining machine-learning models for gamma-ray detection and identification

“…Recently there has been an increase in the use of machine learning to analyze an entire gamma-ray spectrum, often to determine the presence of certain isotopes and quantify their relative strengths [9][10][11][12][13][14][15][16][17][18][19][20][21][22][23][24]. As more complex approaches are used for the detection and quantification of isotopes using gamma-ray spectroscopy, and especially as those approaches are potentially encountered in high-stakes security applications, it will become increasingly necessary for researchers and end-users to understand how algorithms are reaching their conclusions [25].…”