Suppression of surface waves with the ensemble empirical mode decomposition to enhance ultrasound images of reinforced concrete elements

Abstract

The surface waves were suppressed to enhance ultrasound images of reinforced concrete elements with no hard assumption about the surface waves’ time-domain characteristics. The surface waves yield artifacts that degrade image quality and generally hinder the recognition of shallow inclusions and defects. The Ensemble Empirical Mode Decomposition (EEMD) was used to adaptively decompose the signals into their Intrinsic Mode Functions (IMFs). The signals were acquired in a concrete element with rebars at 54, 74, and 94 mm from the concrete element’s surface. The surface waves were generally decomposed to their own IMFs. The IMFs which comprised noise signal components, including the surface waves, were labeled as such based on empirical data. The remaining IMFs were used to construct new signals that generally comprised signal components from reflection at a rebar and the backwall and other low-frequency phenomena. The new signals yielded images in which the surface waves artifacts were not present and the rebars were represented correctly. The Synthetic Aperture Focusing Technique was used for imaging. The method has two main drawbacks. The EEMD performs according to parameters which require extensive ad hoc testing to be selected. The IMFs comprising the surface waves suffered from mode mixing with the backwall at times and hence the backwall was not fully represented in the new images.