Automatic thermograms segmentation, preliminary insight into spilling drop test

The spilling drop test (SDT) is a non-destructive test that uses passive thermography to characterise the property of the building surfaces to absorb and diffuse water. However, the drops made visible by the evaporation process, show unclear contours that do not allow easy quantification of the wetted area. In this work, three image segmentation methods based on different statistical treatments have been considered in order to estimate the wet area. The information acquired by spilling drop test were also compared with standardised methods for the assessment of moisture in porous geomaterials. In addition, the pore size distribution was assessed by nuclear magnetic resonance. The proposed methods were tested by analysing a set of ancient Roman plasters and modern mock-ups to highlight the ability of the thermographic method to provide useful information in the field of conservation of historical building materials.

Automated summary

In this study, three different image segmentation methods were used to estimate the wet area of building surfaces characterized by the spilling drop test. The results were compared with standardized methods for moisture assessment, and pore size distribution was evaluated using nuclear magnetic resonance. The study shows the potential of thermographic methods in providing valuable information for the conservation of historical building materials.