The Study of Cosmic Ray Tomography Using Multiple Scattering of Muons for Imaging of High-Z Materials
Muon tomography is developing as a promising system to detect high-Z (atomic number) material for ensuring homeland security. In the present work, three kinds of spatial locations of materials which are made of aluminum, iron, lead and uranium are simulated with GEANT4 codes, which are horizontal, diagonal and vertical objects, respectively. Two statistical algorithms are used with MATLAB software to reconstruct the image of detected objects, which are the Point of Closet Approach (PoCA) and Maximum Likelihood Scattering-Expectation Maximization iterative algorithm (MLS-EM), respectively. Two analysis methods are used to evaluate the quality of reconstruction image, which are the Receiver Operating Characteristic (ROC) and the localization ROC (LROC) curves, respectively. The reconstructed results show that, compared with PoCA algorithm, MLS-EM can achieve a better image quality in both edge preserving and noise reduction. And according to the analysis of ROC (LROC) curves, it shows that MLS-EM algorithm can discriminate and exclude the presence and location of high-Z object with a high efficiency, which is more flexible with an different EM algorithm employed than prior work. Furthermore the MLS-EM iterative algorithm will be modified and ran in parallel executive way for improving the reconstruction speed.