Focus

Methodologies for a computer aided monitoring of cultural heritage degradation

In order to supply support to Cultural Heritage experts in their work of objective diagnosis of the degradation of historical buildings, in the past decade various specific methodologies were proposed and developed at the IAC. The activities are partially financed by the national project SIINDA and contracts with the Region of Val d'Aosta.
In order to overcome traditional investigation protocols, the research is focussed on the optimization of the attainment of environmental information, chemical-physical data, and images concerning a given monument, along with their analysis, interpretation and presentation. An important constraint comes from the requirement that all proposed tools should be non invasive. The developed methodologies concern data processing and image analysis to extract characteristic features of degradation, the creation of tools which associate chemical and physical conditions to the origin of degradation processes, and the development of mathematical models for the evolution of these processes.
To support as much as possible the expert in monitoring actions, these methodologies are considered as separate modules of an integrated system, which has been designed by analyzing and, subsequently, formalizing the traditional procedures of the monitoring process.
The first phase of pre-monitoring is based on a combination of different components. In fact, the expert's know-how and the methodologies which are traditionally used in conservation protocols, such as sampling of building materials and disposing sensors, both remotely and on the monument, are integrated by techniques of image analysis for colour based segmentation, wavelet analysis and feature extraction for optimal allocation of measure points, tools from artificial intelligence, such as neural networks, for the development of virtual sensors, and mathematical models for the description of the evolution of chemical damage. The mixing, during the pre-monitoring phase, of traditional methods with modern tools from computer science and computational mathematics allows us to collect knowledge about the environment-monument system and to create, at the same time, all necessary tools which will form the framework of a computer aided monitoring system which substitutes the more traditional conservation methodologies.
During the monitoring phase, the virtual sensors and the models for damage evolution, based on only physical and chemical parameter values, will supply information about the conservation state of the monument. Such information is interpreted and translated in alert messages using a knowledge base which is developed a priori, with the help of an expert, for each monument. Moreover, visual and multimodal presentations are created which associate the information with the 3D model of the monument.
All the components of the pre-monitoring phase have been realized and tested on data of a case study (Roman Theatre in Aosta city), while the tuning of the monitoring components is in progress.