PCA-based Method for Managing and Analyzing Single-Spot Analysis Referenced to Spectral Imaging for Artworks Diagnostics

Artworks diagnostics is based on the joint use of several nondestructive techniques to acquire complementary information on the materials. A common practice in the field is to perform the analyses with single-spot analytical techniques, e.g. spectroscopy-based, after a preliminary screening of the artwork with full-field imaging-based techniques. We present a method and its practical implementation for fusing and analyzing data collected using analytical systems that acquire single spot measurements mapped to spectral imaging stacks. The fused dataset of single-spot and imaging observations is analyzed using principal component analysis (PCA). The effectiveness of the method for artworks diagnostics is shown on spectroscopy and imaging datasets of an ancient canvas painting. The results of the PCA analysis on the final fused dataset are compared against the PCA analysis performed on the original datasets from single-spot and imaging measurements taken separately. We propose two practical implementations of the procedure, one based on using graphical user interface (GUI) and open-source GIS software (QGIS), the other one based on an open-source Python module, named SPOLVERRO, specifically developed for this project and released on a public repository. The method allows conservation scientists to analize effectively the heterogeneous datasets acquired in a diagnostic campaign. • single-spot spectroscopy data are referenced on imaging data. • the sampling area of each spectroscopy spot is used for extracting and averaging the respective imaging data values. • the final matrix is analyzed using PCA for extracting further information.