Ultrasonic testing of masonry structures by features extraction and self-organising maps

This paper presents the results of a novel approach to ultrasonic testing of masonry structures. This approach is based on the integrated analysis, in both time and frequency domains, of several features associated with ultrasonic waves propagating through the material. By the Direct Transmission Technique, the received ultrasonic signals have been acquired in a grid of nodes, and several signal features have been extracted. Firstly, each single feature effectiveness in detecting defects has been analyzed. The selected features allow detecting the presence of a cavity within the wall structure under examination. In order to have a more precise definition of the contours of the cavity, a Self-Organising Map has been built taking into account the highdimensional features space. The proposed method has been tested on a full scale stone masonry model with known cavities inside. © 2013 Taylor & Francis Group, London.