N2B1  Instrumentation for Homeland and National Security, Applications of Imaging

Tuesday, Nov. 3  10:30-12:30  Town and Country

Session Chair:  David Wehe, University of Michigan, United States; Michael Wright, Oak Ridge National Laboratory, United States

Show/Hide All Abstracts

(10:30) N2B1-1, Quantitative Evaluation of Volumetric Imaging Systems with Scene Data Fusion

R. Barnowski1,2, A. Haefner2, L. Mihailescu2, K. Vetter1,2

1Nuclear Engineering, University of California, Berkeley, Berkeley, CA, USA
2Applied Nuclear Physics, Lawrence Berkeley National Lab, Berkeley, CA, USA
An approach to gamma-ray imaging dubbed scene data fusion (SDF) has been previously demonstrated, enabling real-time volumetric (i.e. 3D) imaging of gamma-ray source distributions in arbitrarily large and complex environments. SDF is based on merging complementary data streams from auxiliary scene sensors, providing real-time pose tracking and 3D scene reconstruction capability. Gamma-ray imagers can thus be operated in a mobile mode and the tractability of the 3D gamma-ray image reconstruction problem is greatly improved via spatial constraints on the imaging domain provided by the 3D model. While the auxiliary data streams are necessary to enable real time volumetric gamma-ray imaging, uncertainties in these auxiliary data; particularly errors in the pose estimate of the detector, can lead to gamma-ray image artifacts or even failed convergence in the reconstruction. The SDF method has also demonstrated the ability to localize point sources in 3D with very low Compton event data rates (~50 cones/min). This low-count regime is very relevant for real-world source search and mapping applications, therefore quantifying the relevant imaging parameters (resolution, SNR) as a function of measured events is of interest to determine the limitations of SDF. This work explores the impact of uncertainty in the pose and 3D model estimates on image quality as well as the impact of low count rates on reconstruction convergence. A framework has been developed to quantify the accuracy and precision of 3D gamma-ray images for point source measurements. This framework is applied to a set of measurements using strong (mCi) gamma-ray sources as well as simulation results to quantify the impacts of pose uncertainty and low Compton count rates in a way that is generally applicable to volumetric gamma-ray imaging systems. Two distinct imaging systems, an HPGe based Compton camera and a CZT-based handheld imager are used in this work to verify the generality of the uncertainty analysis.

(10:50) N2B1-2, Compton Imaging Data Quality and Imaging Distributed Gamma-Ray Sources

A. Haefner1, K. Vetter1,2

1Applied Nuclear Physics, Lawrence Berkeley National Lab, Berkeley, CA
2Nuclear Engineering, UC Berkeley, Berkeley, CA
In May 2014, our High Efficiency Multimode Imager (HEMI) was flown on a remote control helicopter in Fukushima, Japan for the purpose of mapping Cs-137 contamination. The goal of this measurement is to identify hotspots within a diffuse background of Cs-137. Additionally, the extent of Cs-137 in the trees as compared to the ground is unknown. Imaging in this environment requires excellent imaging contrast, which can be challenging for Compton imaging that has geometric and scattering biases. This has lead to research into Compton imaging data quality. In contrast to point source localization, distributed sources are much more challenging. Some issues that need to be considered include random coincidence events, data sampling, down scattering into the energy of interest, and scattering out of the energy of interest. Our approach to studying these issues is through lab measurements that focus on each issue in an attempt to approach the complexity of the Fukushima mapping environment. We will present methods for analyzing the impact of these issues, along with some possible solutions for overcoming them. For example, sampling issues can be lessened with modified reconstruction algorithms, such as a wavelet regularized ML-EM reconstruction that we have developed for HEMI.

(11:10) N2B1-3, Additional Capabilities of a Compact Neutron Scatter Camera: Active Interrogation, Time-Correlated Pulse-Height Multiplication Measurements, and Gamma Imaging

J. E. M. Goldsmith1, J. S. Brennan1, M. D. Gerling1, P. A. Marleau1, M. Monterial2

1Sandia National Laboratories, Livermore, CA, USA
2University of Michigan, Ann Arbor, MI, USA
Sandia National Laboratories has been developing a mobile fast neutron imaging platform to enhance the capabilities of emergency responders in the localization and characterization of special nuclear material. The mobile imager of neutrons for emergency responders (MINER) is a compact (15” diameter, 36” high, ~90 pounds) neutron scatter camera optimized to provide omni-directional (4-pi) imaging. Although the system performance is tuned for fission energy neutron imaging and spectroscopy, it is also capable of functioning as a Compton camera for gamma imaging. We have previously reported on the fast-neutron imaging and spectroscopy capabilities of this system, and introduced gamma imaging techniques with the same device. Recently, we explored the use of MINER to image actively stimulated highly enriched uranium (HEU) as well as to measure correlated gamma-neutron timing distributions to characterize the multiplication of plutonium containing objects. For the former, a low-neutron-energy americium-lithium (AmLi) neutron source was placed adjacent to the Training Assembly for Critical Safety (TACS) shells, ~25 kg of HEU, stimulating induced fission in the object. For the latter, MINER was operated in an “open array” configuration, placing all sixteen liquid-scintillator detectors at a nearly equal distance from the Beryllium Reflected Plutonium (BeRP) ball, ~4.5 kg of weapons-grade plutonium. Measurements of the timing distribution between correlated pairs of gamma-rays and neutrons emitted by the object enabled us to characterize the neutron multiplication of the object. Gamma imaging provided additional capabilities for localizing both uranium- and plutonium-containing objects. We will present results on all of these capabilities of the MINER system.

(11:30) N2B1-4, Stochastic Image Reconstruction for Non-Proliferation Applications

M. C. Hamel, A. Poitrasson-Rivière, J. K. Polack, S. D. Clarke, M. Flaksa, S. A. Pozzi

University of Michigan, Ann Arbor, MI, USA
The stochastic origin ensembles (SOE) method has been implemented for the reconstruction of neutron and gamma-ray images from a dual-particle imager (DPI). SOE image reconstruction is a stochastic process that creates images by using Markov-chain Monte-Carlo sampling with event origins from the reconstructed backprojection cones. This method has been shown to produce image quality comparable to that of the widely implemented list-mode maximum-likelihood expectation-maximization (MLEM). Previous work has shown that SOE reconstruction is faster than list-mode MLEM. The method is also quickly adaptable to any system configuration because no system response estimate is needed. In nuclear non-proliferation applications, a measurement may contain a small number of source events as well as high background count rates. This study examines the relationship between three important parameters for SOE image reconstruction: the number of measured events, the image pixel size, and the number of iterations necessary to achieve convergence. Cases examined are measurements and Monte Carlo simulations of special nuclear material including mixed-oxide nuclear fuel (category III) and weapons grade plutonium (category I).

(11:50) N2B1-5, Feasibility Demonstration of Two-Dimensional Time-Encoded Fast Neutron Imaging Using a Single Detector Pixel

P. Marleau1, J. Brennan1, M. Gerling1, N. Le Galloudec1, K. McMillan2, A. Nowack1

1Radiation and Nuclear Detection Systems, Sandia National Laboratories, Livermore, CA, USA
2University of California, Los Angeles, Los Angeles, CA, USA
A series of laboratory experiments were undertaken to demonstrate the feasibility of two dimensional time-encoded imaging with a single detector pixel. A prototype two-dimensional time encoded imaging system was designed and constructed. Results from imaging measurements of single and multiple point sources as well as extended source distributions will be presented. Time encoded imaging has proven to be a simple method for achieving high resolution two-dimensional imaging with potential to be used in future arms control and treaty verification applications.

(12:10) N2B1-6, High Speed, Low Dose, Intelligent X-Ray Cargo Inspection

A. Arodzero1,2, S. Boucher1, J. Hartzell1, S. V. Kutsaev1, R. C. Lanza2, V. Palermo3, S. Vinogradov4, V. Ziskin5

1RadiaBeam Technologies, LLC, Santa Monica, CA, USA
2Dept. of Nuclear Science and Engineering, Massachusetts Institute of Technology, Cambridge, MA, USA
3Vertilon Corporation, Westford, MA, USA
4Cockcroft Institute of Accelerator Science and Technology, University of Liverpool, Liverpool, UK
5Physical Science Inc., Andover, MA, USA
The security market requirements for high throughput rail cargo radiography inspection systems include high imaging resolution (better than 5 mm line pair), penetration beyond 400 mm steel equivalent, high scan speeds (>10 km/h, up to 60 km/h), material discrimination (four groups of Z) with 100% image sampling for speed up to 45 km/h for speed up to 45 km/h, low dose and small radiation exclusion zone. To meet and exceed these requirements, research into a new radiography methods and system design, and new detector materials has been initiated. The high speed Adaptive Railroad Cargo Inspection System (ARCIS) is being developed by the team lead by RadiaBeam Technologies. The idea of ARCIS relying on linac-based, adaptive, ramped energy source of packets of short X-ray pulses, new types of fast X-ray detectors, rapid processing of detector signals for intelligent control of the linac, and advanced radiography image processing. The following ARCIS key technologies, which overcomes the limitations of conventional dual energy interlaced cargo inspection systems, will be discussed: - Ultra-fast detection with SiPM readout; - Multi-energy material discrimination in a single scan line which is allows by use of linac based X-ray source of packets of pulses with ramping up energy profile (> 1 MHz energy switching); - Real-time intelligent setting of packet’s maximum energy depend on X-ray attenuation in cargo; - Adaptive dynamic adjustment of the responsivity of each SiPM for increasing effective detector dynamic range; - Scintillation-Cherenkov detector with reduced sensitivity to scatter radiation. Application of these technologies allows maximizing material discrimination, penetration and contrast resolution while simultaneously reducing the dose.