Design and simulation studies on an tomographic imaging detector prototype based on cosmic muons

Cosmic muons are natural radiation sources that have relatively long lifetimes with speeds close to the speed of light. Due to their larger mass they weakly interaction with material and travel longer distances within the materials than X-rays or gamma radiation. Due to these properties cosmic muons are perfect for tomographic imaging. One of the main aspects concerning the concept of the muon tomography is tomographic image reconstruction technique which based on the use of atmospheric cosmic muons are being investigated by various research laboratories around the world. There are various applications for this technique such as nuclear reactor and waste imaging and the recognition of High-Z material in cargo containers for homeland security. There are two well-known image reconstruction algorithms that are widely used in cosmic muon scattering tomography. They are the Point of Closest Approach (POCA) algorithm and the Expectation Maximization-Maximum Likelihood (EM-ML) algorithm both developed by the Los Alamos National Laboratory (LANL) group. The POCA method is a simple geometric algorithm that searches for the closest geometrical point of the incoming and outgoing tracks which is called as the POCA point. The EM-ML method is a statistical algorithm that gives a better result of tomographic image. However, POCA algorithm has weakness in three dimensions space, when the incident and scattered tracks are not coplanar and not intersect at single point and POCA points may occur outside the sample volume. There are much more "background" scattering points which are computation error due to the POCA approximation. The EM-ML method is more difficult to apply and cannot match the requirements of a real time image processing due to the fact that it can no be able to produce results in a small computational time. Generally, in these reconstruction algorithms the standard iterative method, where the probed volume is voxelized, which requires relatively long CPU time. Therefor, these algorithms often does not run in real time and tend to be slow for the homeland security purpose. And there are blurred image especially for lower Z materials that indistinguishable from the background scattering. The present study reports a faster and simple algorithm that is easier to implement and give faster result with smaller error.