Tomographic applications of wavelets in passive nondestructive assay of radioactive waste

An economically reliable and technologically feasible methodology for assaying radioactive waste is essential for properly disposing of newly generated and legacy radioactive waste. The present research has investigated, developed, and integrated tomographic applications of wavelets into the present segmented gamma scanning (SGS) measurement methodology used for passive nondestructive assay of radioactive waste. The SGS measurement methodology was specifically developed to nondestructively measure the radionuclide content within only low-density, homogeneously drum-packaged scrap and waste using the techniques of gamma-ray spectrometry. A result of this research has been the Tomographic Segmented Gamma Scanner (TSGS) system.

This TSGS system supplements a rotation-averaged SGS measurement system with tomographic gamma scanning (TGS) capability that can specifically take into account the effects of larger density and radionuclide distribution non-uniformities when present. This TSGS system allows the throughput to remain high with a less detailed but faster rotation averaged segment measurement of homogenous radioactive waste, while the TGS capability provides a more detailed but slower spatial radionuclide measurement for assaying heterogeneous radionuclide waste as necessary. Within the experimental part of this research, (1) the tomographic experimental feasibility was established, (2) a TSGS system integrated calibration was designed, (3) collimation sampling problems were studied, (4) SGS and TGS modality-specific comparisons were shown, and finally, (5) the TSGS system precision was measured.

An investigation of the applications of wavelets in computerized tomography has lead to the development of a radionuclide computerized tomography (RCT) algorithm able to accurately calculate the measured radionuclide content within heterogeneously packaged scrap and waste. Within the computational part of this research, (1) the optimization of a relaxation parameter that determines the iterative convergence rate within the RCT algorithm was investigated, (2) an efficient wavelet-interpolation method for refining the tomographic reconstruction resolution to allow accurate forward and back-projecting interpolation using simple nearest-neighbor grid-point assignment was developed, (3) the multiresolution analysis of tomographic image reconstructions with applications in wavelet filtering was studied, (4) an exploration of how Daubechies wavelets influence tomographic reconstructions was investigated, and finally, (5) simulated tomographic profile projection data using the Monte Carlo N-Particle Transport Code System (MCNP) was successfully benchmarked against experimental data.