IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 56, NO. 2, APRIL 2009 479 Standoff 3D Gamma-Ray Imaging Lucian Mihailescu, Kai Vetter, and Daniel Chivers Abstract—We present a new standoff imaging technique able to provide3-dimensional (3D)imagesof gamma-ray sources distributed in the environment. Unlike standard 3D tomographic methods, this technique does not require the radioactive sources to be bounded within a predefined physical space. In the present implementation, the gamma-ray imaging system is based on two large planar HPGe double sided segmented detectors, which are used in a Compton camera configuration. A LIDAR system is used in conjunction with the gamma-ray imaging system to confine the gamma-ray image space to the interior of physical objects situated within the detection range of the gamma-ray imager. This approach results in superior image contrast and efficient image reconstruction. Results demonstrating the operating principle are reported. Index Terms—Compton scatter imaging, gamma-radiation, ra- diation imaging, semiconductor detectors. I. I NTRODUCTION T ODAY, imaging of radioactive gamma-ray sources in 3 di- mensions (3D) is mainly the subject of Positron Emission Tomography (PET) and Single Photon Emission Computer To- mography (SPECT) techniques [1]. These imaging techniques require the radioactive tracer of interest to be distributed within a predefined bounded physical space, also called image space. For accurate imaging, the detection system also requires physical access around the image space. These constraints are acceptable for many applications, especially in medical imaging, where the subjects of interest are known, and can be contained within a predefined space. However, there is an increased interest from the homeland security and nuclear nonproliferation community for a gamma-ray imaging system able to provide 3D images of radioactive materials distributed in a physical space of arbitrary extent. A generic scenario assumes multiple radioactive isotopes heterogeneously distributed from near field distances to far field distances (centimeters to tens of meters) around the imaging sensor. Applications range from nuclear safeguards, such as mapping the distribution of special nuclear materials in facili- ties across the nuclear fuel cycle, to homeland security, such as Manuscript received June 30, 2008; revised November 04, 2008 and January 09, 2009. Current version published April 08, 2009. This work was supported by the Domestic Nuclear Detection Office of the Department of Homeland Security and by the NA-24 Office of the National Nuclear Security Administration. L. Mihailescu is with the Lawrence Berkeley National Laboratory, Berkeley, CA 94720 USA (e-mail: lmihailescu@lbl.gov). K. Vetter is with the Lawrence Berkeley National Laboratory, Berkeley, CA 94720 USA, and also with the Nuclear Engineering Department, University of California, Berkeley, Berkeley, CA 94720 USA (e-mail: kvetter@nuc.berkeley. edu). D. Chivers is with the Nuclear Engineering Department, University of Cali- fornia, Berkeley, Berkeley, CA 94720 USA (e-mail: chivers@berkeley.edu). Color versions of one or more of the figures in this paper are available online at http://ieeexplore.ieee.org. Digital Object Identifier 10.1109/TNS.2009.2015304 stand-off detection of radioactive sources in the environme or search and surveillance of nuclear materials. Since there is no predefined bounded image space, other than the detection range of the gamma-ray sensor, these applications require ferentsolution than the one offered by standard tomograph methods. The detection range of the imager is determined experimental conditions, including, but not limited to: air a uation, presence of attenuating obstacles, and relative stre of background radioactivity. Even though the detection ran not a predefined value, it depends on the purpose of the m surement, and acts as a first bounding factor in our approach to gamma-ray imaging. Therefore, the image space can be sumed to extend to the volume of a sphere around the ima system, with a radius equal to a preset detection range. Ho the discretization of the whole space within the detection r leads to a huge image matrix, which makes image reconstr tion a very computing intensive task. In this paper, we pre a methodology by which the gamma-ray image space is sig cantly reduced by using the gamma-ray imager in combina with a range scanner that confines the image space to the rior of physical objects situated within the detection range gamma-ray imager. In the following section, a description imaging method will be given, with an emphasis on the ori method used to build the image space, and on the calculat the backprojection weights. Afterwards, the instruments u the demonstration measurements will be described, and fi the results of a preliminary demonstration measurement w presented with comments on its performance. II. S TAND -O FF3D G AMMA -R AY I MAGING Imaging ofgamma-radioactive isotopes distributed in the environment has been previously done predominantly using collimator-based imagers such as coded aperture cameras Using such techniques, two-dimensional (2D) mapping of sources from long-range distances has been demonstrated In the present work we aim to extend long-range (or stand-off) imaging to three dimensions. Standard tomographic methods may not be suitable to provide a practical solution for stan 3D gamma-ray imaging because of two main reasons: 1) n mally,no physical access is available for the imaging system around the radioactive sources; and 2), the existing tomog reconstruction algorithms do not provide a solution for the of an arbitrary image space. When the image space is not pre-defined, bounds can be artificially imposed by taking into account the expected detection range of the system. Even however, a large number of image elements (pixels or vox will result.Because ofthis,image reconstruction becomes a very computing intensive process, which is prohibitive for many applications. 0018-9499/$25.00 © 2009 IEEE