X-Ray Diffraction Imaging
R4-C.2

Download Project Report

Project Description

Overview and Significance

This project investigates the development of improved automated explosives detection and classification algorithms through the fusion of multiple modalities. Of particular interest are techniques that can potentially penetrate luggage and complement the information provided by dual-energy X-ray imaging. Our effort is focused on extracting additional signatures from X-ray excitation beyond the conventional density and effective atomic number by using X-ray diffraction. X-ray diffraction can provide information concerning coherent scatter distributions at different locations for different momentum transfer levels. This coherent scatter distribution is a function of the electron distribution in molecules and provides a surrogate signature that can serve to specify the type of material in a manner that is complementary to the typical dual-energy absorption profiles.

There are major limitations in current X-ray diffraction systems. First, there is a need to localize the coherent scatter to regions so that the signatures can be associated with specific objects inside luggage. Second, the resulting scattered signals from different volumes inside the luggage undergo complex absorption and secondary scatter on the way to detectors, which must be compensated for. Third, the measured signals are relatively weak, as the fraction of scattered photons is spread over volumetric angles in a frequency-dependent manner, so the signals collected by each detector are limited. In this project, we investigate different algorithmic and architectural approaches that can combine information from multiple frequencies and scattering angles at the image formation stage, leading to improved signal-to-noise ratio (SNR) and subsequently improving threat detection and classification. Our algorithms provide the basis for coherent scatter image reconstruction for future luggage inspection systems, providing information beyond density and effective atomic number. Our architecture comparisons highlight the relative performance of alternative architectures for practical X-ray diffraction imaging systems.

X-ray diffraction imaging (XDI) is currently an emerging technology that synthesizes two important characteristics of X-rays: their ability to form images and the ability to perform material analysis via representative X-ray diffraction profiles.
Phase 2 Year 2 Annual Report
Project Leader
  • David Castañón
    Professor
    Boston University
    Email

Faculty and Staff Currently Involved in Project
  • W. Clem Karl
    Professor
    Boston University
    Email

Students Currently Involved in Project
  • Parisa Babahedarian
    BU