Computation of image features for Full-wavefield Characterization in fast non-contact inspections

Ultrasonic wavefield imaging techniques are typically applied on full acoustic wavefield data acquired over the area of the structure to be inspected. Such techniques can be very useful, since they can be implemented with non-contact sensors and are capable to etect and locate damages, but present also some limitations, including slow data acquisition and lack of accuracy. This research addresses both these challenges and presents a fast and robust non-contact wavefield imaging method. The proposed method adopts the Compressive Sensing approach as a mean to speed up the acquisition process through a random spatial undersampling and computes multiple image features to perform the characterization of the inspected medium. Several experiments were performed on aluminium and epoxy structures to illustrate pros and cons of the different damage features. The results illustrate that the combined use of wavefield imaging and compressive sensing provides sufficient information to non-destructively evaluate a specimen in many situations of interest.