An automated sonic tomography system for the inspection of historical masonry walls

Abstract

Background: The conservation of the built masonry heritage requires a comprehensive understanding of its geometrical, structural, and material characteristics. Non-destructive techniques are a preferred approach to survey historical buildings, given the cultural value of their fabric. However, currently available techniques are typically operated manually, consuming much time at operational and processing level and thus hindering their use for the on-site inspection of heritage structures. Methods: A novel automated sonic tomography system was designed and built to inspect and obtain information about the inner structure and damage of historic masonry walls. The system consists of a hitting device mounted on a frame that can be placed adjacent to the wall under analysis. The hitting device can move along the surface within the frame area in X, Y and Z directions, generating the sonic wave. The receiving system is a scanning laser vibrometer, able to measure from the distance the displacement of a focused point over time, recording the wave when it reaches the opposite surface. Results: Six stone masonry walls with different interior geometries were constructed at the laboratory by a professional stonemason. The construction of the walls was carefully documented, including the generation of detailed photogrammetric models of each single stone. The system was applied to survey the six masonry walls. Since the inner morphology of the walls is known, the resulting tomographic images could be compared with the ground truth. Conclusions: Automating the inspection allowed to collect thousands of data in a few hours. New software was also developed to automate the processing of the data. Results are expected to highlight the potential of tomography to obtain quantitative information about the interior of heritage structures, while providing new tools that make the implementation of the technique more practical for professionals. Data, software and models have been made publicly available.