Stacked Elasticity Imaging Approach for Visualizing Defects in the Presence of Background Inhomogeneity

The ability to detect spatially-distributed defects and material changes over time is a central theme in structural health monitoring. In recent years, numerous computational approaches using electrical, electromagnetic, thermal, acoustic, optical, displacement, and other nondestructive measurements as input data for inverse imaging regimes have aimed to localize damage as a function of space and time. Often, these regimes aim to reconstruct images based off one set of data disregarding prior information from previous structural states. In this paper, we propose a stacked approach for one increasingly popular modality in structural health monitoring, namely quasi-static elasticity imaging. The proposed approach aims to simultaneously reconstruct spatial changes in elastic properties based on data from before and after the occurrence of damage in the presence of an inhomogeneous background. We conduct numerical studies, investigating in-plane plate stretching and bending, considering geometries with various damage levels. Results demonstrate the feasibility of the proposed imaging approach, indicating that the inclusion of prior information from multiple states visually improves reconstruction quality and decreases root mean-square error (RMSE) with respect to true images.