From Event: SPIE Optical Metrology, 2019
The analysis of painting materials and techniques provides important information to history and archaeology research, as well as to scientists and conservators involved in the conservation of objects of cultural heritage. However, the examination of statistically significant number of objects is required to understand the material use typical of a class of objects (e.g. paintings from a certain geographic region, in a certain period, for a certain purpose). This necessitates efficient data collection and the development of methods for their effective analysis. The spectral imaging system developed in our group, PRISMS, enables automated high resolution spectral imaging in the visible/near infrared regime of paintings of any size. The automatic clustering of the spectral reflectance data at the pixel-level can be used for the initial classification of the areas according to their reflectance spectra. In this study, a new clustering algorithm, based on self-organised mapping (SOM), for the sequential automatic analysis of large spectral datasets is presented. The development of such methods makes the analytical procedure more efficient by reducing the number of areas that need to be further examined based on their unique reflectance spectra. The preliminary clustering of the spectral imaging data is supported by the high-resolution x-ray fluorescence (XRF) maps. The multimodal spectral characterisation of the painting materials is completed with the application of high spectral resolution Fibre Optic Reflectance Spectroscopy (FORS) and Raman spectroscopy on each cluster area. Optical coherence tomography (OCT) is used to examine the 3D microstructure of the substrates as well as the painting technique through the examination of the stratigraphy. Examples based on data from collections of Peruvian paintings on paper substrate, Chinese export paintings and Peruvian paintings made in the style of Chinese export painting will be presented.
© (2019) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Sotiria Kogou, Lynn Lee, Golnaz Shahtahmassebi, and Haida Liang, "A novel methodology for the automatic analysis of large collections of paintings (Conference Presentation)," Proc. SPIE 11058, Optics for Arts, Architecture, and Archaeology VII, 110580Q (Presented at SPIE Optical Metrology: June 25, 2019; Published: 22 July 2019); https://doi.org/10.1117/12.2527611.6062676747001.