Computer-aided monitoring of buildings of historical importance based on color

This paper presents an application of image processing tools in the field of cultural heritage. The proposed strategy allows us to improve and make semi-automatic the study of chemical decay causing visible changes in color of some regions. So, it is possible to semi-automatically detect the regions corrupted by a specific kind of decay from a color image showing the degraded surface of a building of historical importance. The decay will be indicated by the expert via mouse clicks on the image under study. The second phase of the proposed framework consists of selecting a subset of the regions achieved in the first one. The number of regions in the subset is selected by the expert. The remaining operations are automatically performed via an optimization approach. Both phases try to embed experts' knowledge. The final result will be a map of points on the surface from which extracting a bit of material to study in depth via chemical laboratory tests. We outline the potentialities of the proposed non-invasive technique, oriented to avoid some drawbacks of the classical “naked-eye” approach. Even though the proposed framework is guided by the expert, it is semi-automatic and able to reduce typical human limits such as subjective and emotive state, visual system performances and so on. Experimental results, achieved on some images of the Roman Theatre in Aosta, will be presented and discussed.