Abstract
A collimator is a very important part of a gamma camera. There are many types of collimator, but the most common collimator used for general type of imaging is a multihole collimator. Despite its good resolution, the multihole collimator is heavy and bulky, therefore it is difficult to handle. In order to tackle this problem, this thesis introduces a practical model of a lighter wire-mesh collimator and finds the optimal configuration for it. The gamma camera and the collimator are simulated using the MCNPX code. First, we simulate the multihole collimator gamma camera and use the results from these experiments as the standard images to be compared with those produced with the wire-mesh collimator. The sources that we used are a point source, a planar square source and two point sources, and for all the cases, photons at 140keV are simulated. We show that the proposed hardware design in conjunction with image processing can produce images of comparable quality with those of the full collimator with only 39.5% of its weight. The image processing method we use is based on Wiener filtering. Wiener filtering relies on the estimation of the power spectra of the noise and the ideal image. The analysis on the estimation of power spectra is also included in this thesis, and we showed that the combination of the Goodman and Belsher technique with a constant, is the optimum method to be used with the Wiener filtering technique. Key words: Wire-mesh Collimator, Multihole Collimator, Gamma Camera, Wiener Filter.