Detailed segmentation for the analysis of rubble masonry substrates is complex due to the lack of uniformity in size, geometry, coursing, bonding, and materials composition and texture of the individual stones. State-of-the-art technologies, such as Terrestrial Laser Scanning (TLS) or photogrammetry, deliver precise geometrical and coloured data that can be processed by means of innovative techniques to segment and label these masonry units. The automatic segmentation of individual stones supports further analysis of rubble masonry via several parameters related to their geometry and face colour. These can be utilised for deciphering architectural construction methods and changes in materials, as well as being employed for practical maintenance and repair operations. This paper presents a new strategy for investigating historic masonry substrates, especially from the perspective of data processing methods supporting the automated segmentation and labelling of rubble stone walls. In particular, an innovative method based on the Continuous Wavelet Transform (CWT) is proposed for the automatic segmentation of rubble stone walls, which subsequently enables the automated analysis of the individual stones and mortar regions. Different experiments have been conducted on two significant Cultural Heritage (CH) buildings in Scotland, with the results clearly demonstrating the potential of the proposed method for historic interpretation and analysis.
|11th International Conference on Structural Analysis of Historical Constructions
|11/09/18 → 13/09/18