Scatter Correction and Contrast Improvement of Concrete Objects Using Monte Carlo Method

Abstract

Estimating the load-carrying capacity of concrete structures is a major challenge as the infrastructure industry thrives to reduce the consumption of concrete for environmental concerns as well as improving the cost and time efficiency of infrastructural projects. Quantifying such load-carrying capacity needs a detailed study of the micro-structures of the concrete mixture composed of stones, reinforcement, and cement matrix. Computed Tomography (CT) provides a reliable nondestructive technique to analyze and segment the different materials within the concrete structures. However, CT is normally accompanied by different kinds of artifacts that highly degrades the quality of the reconstructed image. Among these artifacts, scatter imposes severe degradation of the quality of the image including reduced contrast between different materials, cupping and streaks artifacts. Several methods were used to remove the scatter effect. The most effective way to correct for the scatter effect is by the use of the Monte Carlo (MC) method due to the accurate modeling of the physics accompanying the photon matter interaction. In this work, a fast MC photon transport model accelerated over the use of the GPU(s) platforms is developed. It is extensively verified against several well-known simulators and experimentally has been used to correct the scatter effect and to improve the contrast between the concrete component.