GPU-accelerated Simulator for Optical Tomography applied to Two-Phase Flows

Abstract

Optical tomography (OT) is a modality of tomographic imaging that can provide cross-sectional imaging of phase distributions in two-phase pipe flows, thus having potential application in process monitoring. Due to the strong effect of refraction and reflection on phase boundaries (e.g. gas-liquid interfaces), OT measurements of two-phase flows cannot be modeled using hard-field assumptions. This fact renders the use of traditional tomographic reconstruction techniques unsuitable for this problem. In this paper, we present a GPU-accelerated system capable of providing simulation of light transport in an OT imaging system. Sinograms of real OT measurements taken of a phantom are visually compared with simulations, yielding positive results. Quantitative validation of the simulator is left for future works.