Improvements of the Multilayer Holo-Acoustic Tomography by Iterative and Algebraic Techniques

Abstract

The scanning tomographic acoustic microscopy (STAM) using the holographic principle has been recently proposed and the algorithm known as “back-and-forth propagation”(BFP) has been devised for the reconstruction of multi-planar objects. The BFP algorithm is a holographic reconstruction coupled with a tomographic processing of multiple data set obtained by the variation of insonification frequency or transducer angle. The tomographic processing is a superposition process applied to the holographically reconstruction images based on the principle that the structures on the reconstruction plane is insonification-frequency or angle invariant, while the diffraction noise (blurred image) coming from out-of-reconstruction planes is frequency or angle dependent. In practice, often the cancellation is incomplete and reconstructed image is blurred by the diffraction from other image planes. In this paper, two methods of restorations, namely iterative and algebraic restorations, have been proposed with which diffraction noise can be largely removed. Simulation results of the iterative method show substantial improvements in image quality by only a few iteration steps. The algebraic method which solves the simultaneous equations of the coefficients related with the propagation transfer function has better computational efficiency than the iterative algorithm. The algebraic technique provides accurate restoration if a noise free environment is available. On the other hand, the algebraic method appears to be more susceptible to random noise. Study of these two methods by computer simulations is performed and results are shown to demonstrate the efficacy of the two algorithms when they are applied to the BFP algorithm.