Backscatter Imaging: Obtaining Images at Multiple Depths From a Single View
B Juneja1*, F Bova2, E Dugan3, D Shedlock4, (1) ,Gainesville, FL, (2) Univ Florida, Gainesville, FL, (3) Nucsafe, Oak Ridge, TN, (4) Nucsafe, Oak Ridge, TNTH-A-213CD-11 Thursday 8:00:00 AM - 9:55:00 AM Room: 213CD
Purpose: To demonstrate that by changing cut-off planes of a Scatter X-ray Imaging (SXI) system and that through successive image subtraction, the depth and density information about an object can be derived from a single view.
Methods: Phantoms of various thickness and complexity were simulated in a Monte Carlo based transport code MCNPx. An analytic mathematical model of the system was created in MATLAB to verify the initial MCNP results. When necessary, MCNP ptrac files were extracted using a code written in MATLAB for details of all photon interactions. To verify the proposed method, an air cavity and a piece of bone were simulated in a water phantom. Multiple simulations were performed by varying the collimation length and depth of the objects inside water. Initially, only the central axis signal from the phantom was considered, and later a series of backscatter radiographs were obtained using MCNP to validate the method.
Results: It was shown that introducing bone and air into a phantom reduced the detector signal. Collimating to the surface plane eliminated a plethora of surface scatter. The near surface slices contributed the most scatter signal, which attenuated exponentially deeper in the phantom. Image subtraction of multiple collimated cut-off planes showed an expected 60% and 100% drop in signal when bone and air, respectively, were present. However, in successive images, the water signal was misestimated due to attenuation by the object above. This was reduced, in both cases to less than 5%, by applying attenuation corrections for these images.
Conclusions: Monte Carlo simulations verify the possibility of obtaining depth and density information of objects in a phantom from a single view using SXI systems if necessary attenuation corrections were applied. There is a need for a sophisticated iterative algorithm that assess these attenuation corrections for each slice.