Linear Reconstruction Methods for Large Thick Aperture Imaging

Abstract

Large thick aperture imaging method is proposed to measure the radiation intensity distribution of radiation source whose size is several centimetres. The new method contains two steps which are coded imaging and image reconstruction. Geant4 is used to simulate the coded imaging process. Wiener filter, Lucy-Richardson, Regularized filter, blind deconvolution and CT reconstruction method are realized using Matlab software. Wiener filter and Regularized filter produce artifacts. CT reconstruction method can’t reveal the right size of 60Co-γ source because of the overlapping of knife edge images. Both L-R and blind deconvolution are effective. If the aperture is moved 0 mm, 2 mm, 4 mm, 6 mm along x-axis, the reconstruction result using L-R method has a larger correlation coefficient with source image. However, L-R method gets bad result if the aperture rotates more than 1.4°. Under this condition, blind deconvolution is the better choice.