Login
TitleAdvanced Data Processing for Volumetric Computed Tomography Data (Masters Thesis)
Author(s) Matthew J. Sheats
Year September 2000
SchoolCenter for Image Processing and Integrated Computing, University of California, Davis
Download
BibTeX
Abstract Computed tomography has long been used by industry as a non-destructive evaluation technique for discovering flaws in parts before their use. Traditional linear array computed tomography systems acquired data at a rate that could be reconstructed simultaneously with data acquisition. With new area detectors, the data rate of acquisition has increased exponentially, and the reconstruction algorithm does not scale linearly with increased data. FlashCT DPS (Flat Panel Amorphous Silicon High Resolution Computed Tomography, Data Processing System) is a software package developed at Los Alamos National Laboratory to meet the data processing demands of new flat panel array computed tomography. FlashCT DPS combines several features unique to industrial computed tomography systems. It addresses traditional usability problems by providing an intuitive graphical user interface and powerful analysis tools. It includes the three major computed tomography reconstruction algorithms: parallel backprojection, fan beam resorting, and Feldkamp cone beam. It also provides visualization tools for examination of data after processing. Finally, it is able to perform distributed data reconstruction with a near linear speed increase as a function of the number of processors used. Additionally, utility programs have been developed to support project editing and computing cluster management for the FlashCT system. FlashCT DPS runs on a standard PC platform and operates well on a variety of processor and memory configurations.