PARALLEL BEAM NEUTRON COMPUTED TOMOGRAPHY RECONSTRUCTION WITH VARIOUS FILTERS

Abstract

PARALLEL BEAM NEUTRON COMPUTED TOMOGRAPHY RECONSTRUCTION WITH VARIOUS FILTERS. This paper presents a simple calculation using the MatLab application code for five different filter functions to obtain the best image quality for parallel beam neutron computed tomography. Shepp-Logan, Ramachandran- Lakshminarayanan, Cosine, Hamming and Hann Filter and without filter were used for a cross-section of a car coil ignitor reconstruction. The reconstruction was also done by varying the number of projections namely 180, 90, 45 and 12. The quality of the image can be determined by viewing the image directly or by using grey level curve to show the signal to noise image. The good images were determined by their signal-to-noise ratio (SNR) value. The high SNR values were obtained from the images which were reconstructed by using hamming and Hann filter. On the other hand the lowest SNR was obtained when the image was reconstructed without any filter. The number of the projections were also influenced the image quality, the more the projection number is used the better the image quality is obtained.