A Hardware and Software System for Tomographic Research: Reconstruction via Regularization

Abstract

Tomographic reconstruction by integral means preserves the prior information of a reconstructed image. It is difficult to reconstruct high-quality CT images by integral means if the measured parameters or object properties do not satisfy the requirements of the model used in reconstruction (few angles of rotation, angles that are not homogeneously distributed, the object containing regions critically different in absorption at the probing energy). Algebraic approaches are time-consuming. However, they allow consideration of the prior information of a setup or image. Results are prevented from a laboratory tomographic experiment on a test object. A highly absorbing inclusion is placed inside the object. A polychromatic beam is used. Results obtained using integral and algebraic reconstructions are compared. The reconstruction procedure is described in detail for the latter. The algebraic approach uses information on the presence of a highly absorbing region.