skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Nuclear tomography using multiple pinhole imaging with inversion reconstruction

Miscellaneous ·
OSTI ID:6321627

This dissertation describes the development of a tomographic technique for the imaging and reconstruction of extended sources. The development begins with the vector matrix representation of a coded aperture imaging system, and an inversion reconstruction approach. This initial system, based on a planar aperture design and a multi-planar representation of an object, exhibited certain problems related to the rapid degeneration of system conditioning with increase in object size, and inadequacy of the multi-planar model when applied to extended sources. Conditioning problems are reduced by switching from a planar to a linear aperture. Further improvement is obtained by the implementation of time coding, which effectively increases the number of independent equations forming the system matrix. An extended source model is implemented by the use of volume elements (voxels) rather than planar elements (pixels) to represent the object space. Additionally, to increase the accuracy of the system model, photon attenuation and solid angle attenuation models are used, and an ideal and a radius model of the aperture pinholes are explored. During the design phase of the study, the pinhole radius model was found to give more accurate reconstructions than the ideal pinhole model. The pinhole radius model was therefore used for the remainder of the study. A system was designed to reconstruct an oval-shaped phantom, and three homogeneous sources were made for testing and optimization. Modifications to the system geometry were tested, using simulations incorporating Poisson noise statistics to test the conditioning of the resulting system matrices. The accuracy of the system models were tested by simulation and experiment, using the homogeneous sources. Finally, reconstructions were obtained for the oval phantom.

Research Organization:
Toledo Univ., OH (USA)
OSTI ID:
6321627
Resource Relation:
Other Information: Thesis (Ph. D.)
Country of Publication:
United States
Language:
English