3082 IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, VOL. 60, NO. 9, SEPTEMBER 2011 Three-Dimensional Nondestructive “Sampling” of Art Objects Using Acoustic Microscopy and Time–Frequency Analysis Georgios Karagiannis, Dimitrios S. Alexiadis, Student Member, IEEE, Argirios Damtsios, George D. Sergiadis, Member, IEEE, and Christos Salpistis Abstract—The microsampling destructions, which are caused by the sampling procedures of analytical spectroscopic methods, are, in most cases, not permitted to art objects, which are ex- tremely valuable, rare, and fragile. Consequently, the develop- ment of nondestructive analysis techniques becomes a necessity. In this paper, we present a technique and method for the nonde- structive identification of the stratigraphic structure of the paint layers of art objects. Using acoustic microscopy, in combination with time–frequency representations, the continuous or discrete wavelet transform, or the Hilbert–Huang transform, the depth profile of the stratigraphy is determined. Index Terms—Acoustic microscopy, art objects, nondestructive testing (NDT), time–frequency (TF) analysis, ultrasounds, wavelet transform (WT). I. I NTRODUCTION T HE STUDY of the art objects’ paint layer stratigraphy is crucial for their documentation since it provides im- portant information related to the painting technique and to previous restoration attempts. Analytical spectroscopic meth- ods, which require a microsampling operation, applied in conservation science were well established during the last 20 years [1]. However, the nondestructive analysis of art objects becomes a necessity, because the microdestructions caused by the sampling procedure are, in most of the cases, not permitted to the objects under study, which are extremely valuable, rare, and fragile. The nondestructive analysis tech- niques are generally not so effective, compared with the an- alytical microsampling techniques. However, the support of these methods with soft-computing techniques can minimize their ambiguity. In this paper, we propose an acoustic micro- Manuscript received October 12, 2010; revised January 17, 2011; accepted January 28, 2011. Date of publication April 19, 2011; date of current version August 10, 2011. This work was supported by the CHARISMA and InfrArtSonic Projects, which are funded by the European Commission. The Associate Editor coordinating the review process for this paper was Dr. Robert Gao. G. Karagiannis, G. D. Sergiadis, and C. Salpistis are with Aristotle Univer- sity of Thessaloniki, 54636 Thessaloniki, Greece. D. S. Alexiadis is with the Department of Electronics, Technical Education Institute, 57400 Thessaloniki, Greece. A. Damtsios is with Ormylia Foundation, 63071 Ormylia, Greece. Color versions of one or more of the figures in this paper are available online at http://ieeexplore.ieee.org. Digital Object Identifier 10.1109/TIM.2011.2124730 scope supported by a set of soft-computing methods, which are based on time–frequency (TF) signal processing techniques for the nondestructive depth profiling of the stratigraphies of art objects. Up to now, there is a lot of work done in the field of artworks’ documentation with respect to material identifica- tion [1], [2]. The most widely used analytical methods are given here. 1) Multispectral imaging from the visible up to the near in- frared area of the spectrum, providing surface information and information from the underlayers. The method is non- destructive since no contact with the artwork is required, but it provides only qualitative and not quantitative results [4]–[7]. 2) Ellipsometry, providing mainly surface information and also information from the underlayers, in the case that they are optically transparent. No information is provided for pigment materials, which are not optically transpar- ent. This method is nondestructive, but it provides only qualitative results [8], [9]. 3) Unilateral nuclear magnetic resonance: The information provided by this method is mainly associated with the structural stability and not with pigment identification. The method is nondestructive since no contact with the artwork is required. It is still under research but very promising to provide rich qualitative and quantitative results [10], [11]. 4) X-ray fluorescence, providing mainly elemental analysis information from the surface, as well as information from the under-layers. No depth profiling information is provided since it is not known from which layer the information is received. The method is nondestructive, and it provides qualitative results. The possibility to acquire quantitative results is under re- search [12]. 5) Raman, providing mainly information related to the sur- face information and to inorganic materials. In order to acquire information from the depth profile, usually microsampling is required. It provides only qualitative results, while the extraction of quantitative results is under research. 6) Fourier-transform infrared spectrometry: The informa- tion it provides is mainly related to the surface and to 0018-9456/$26.00 © 2011 IEEE