Mathematical osmosis imaging for multi-modal and multi-spectral applications in Cultural Heritage conservation

TL;DR

This paper introduces a dual-mode mid-infrared imaging workflow utilizing osmosis PDE for detecting subsurface damages in frescoes, improving image quality and data integration in Cultural Heritage conservation.

Contribution

It presents an efficient numerical approach for osmosis PDE on large images, enhancing damage detection and multispectral data integration in cultural heritage applications.

Findings

Effective detection of sub-superficial voids in frescoes

Improved image contrast and detail preservation

Enhanced multispectral data integration

Abstract

In this work we present a dual-mode mid-infrared workflow [6], for detecting sub-superficial mural damages in frescoes artworks. Due to the large nature of frescoes, multiple thermal images are recorded. Thus, the experimental setup may introduce measurements errors, seen as inter-frame changes in the image contrast, after mosaicking. An approach to lowering errors is to post-process the mosaic [10] via osmosis partial differential equation (PDE) [12, 13], which preserves details, mass and balance the lights: efficient numerical study for osmosis on large images is proposed [2, 11], based on operator splitting [8]. Our range of Cultural Heritage applications include the detection of sub-superficial voids in Monocromo (L. Da Vinci, Castello Sforzesco, Milan) [5], the light-balance for multi-spectral imaging and the data integration on the Archimedes Palimpsest [10].