LoboVault Home

A prototype Compton imager : simulations, measurements and algorithm development

LoboVault

Please use this identifier to cite or link to this item: http://hdl.handle.net/1928/10302

A prototype Compton imager : simulations, measurements and algorithm development

Show full item record

Title: A prototype Compton imager : simulations, measurements and algorithm development
Author: Tornga, Shawn
Advisor(s): Cooper, Gary
Committee Member(s): Hecht, Adam
Sullivan, John
Department: University of New Mexico. Dept. of Chemical and Nuclear Engineering
Subject: Compton
Imaging
LC Subject(s): Gamma ray detectors.
Compton effect.
Imaging systems--Data procesing.
Expectation-maximization algorithms.
Degree Level: Masters
Abstract: Compton imaging is a gamma ray imaging technique that has possible applications in the nuclear and medical industries. Compton imaging can localize the origin of a scattered gamma ray, to a cone, using a minimum of two position and energy measurements. The projection of many cones will overlap at a common point leading to the actual source position. A prototype Compton imager (PCI) has been constructed at Los Alamos National Laboratory (LANL) that uses a combination of silicon and CsI detectors. A model of the PCI has been simulated and validated, showing good agreement with measured data. The angular resolution of the PCI was measured to be 0.156 radians FWHM. Additionally, a method for determining the source-to-detector distance in the near field has been developed and demonstrated. The algorithm presented has the ability to determine the source-to-detector distance in the near field within 1 cm of the actual distance. While traditional back-projection algorithms are adequate for imaging of single point-like sources they are not sufficient to resolve extended shapes or closely spaced multiple point-like sources. Iterative algorithms may provide the necessary deconvolution. Maximum Likelihood Expectation Maximization (MLEM) is an iterative statistical algorithm that reconstructs the most probable source distribution for a given data set. Normally, MLEM makes computations for each possible combination of energy and position, however this becomes a prohibitively large problem for both analysis time and hardware memory limits, depending on the data set. List Mode MLEM attempts to circumvent the calculation of every combination of energy and position and relies on the probability of the given data being observed. An algorithm using List-Mode MLEM is of interest because the number of calculations required for reconstruction is substantially less than that of other iterative processes but will enable imaging of both point and extended sources. This type of algorithm has been written and successfully applied to both experimental data from the PCI and simulations of the PCI. The algorithm will be demonstrated to improve detection of both point-like and extended sources.
Graduation Date: December 2009
URI: http://hdl.handle.net/1928/10302


Files in this item

Files Size Format View
thesis_final_tornga.pdf 9.463Mb PDF View/Open

This item appears in the following Collection(s)

Show full item record

UNM Libraries

Search LoboVault


Browse

My Account