PROCESSING AND OPTIMIZATION TECHNIQUES SPECTRA RAMAN THROUGH FUZZY LOGIC: APPLICATION TO THE IDENTIFICATION OF MATERIAL PICTORIALSummary: The investigation of the materials constituting artwork, which falls this thesis, is one of the applications of Raman spectroscopy has been growing in importance in recent decades. The successful implementation of this technique in the field of art is that it is one of the analytical techniques that, for their characteristics no-destructividad and feasibility analysis situ best suited to the demands of this kind of study. Thus, through this analysis, we can get to identify materials that constitute a work, which is vital both for their dating, authentication and documentation, to be able to perform rigorous conservation and restoration. The Raman spectroscopy is a technique that photonics is based on the analysis of the tiny fraction of light scattered by the sample analyzed the influence it a beam of monochromatic light, the scattered light contains some changes frecuenciales (Raman bands) with respect to the incident light they are characteristic of the molecular composition of the material analyzed. Specifically, the identification of a pictorial materials using Raman spectroscopy is achieved by locating the position frequency bands present in its Raman spectrum. The set of Raman bands and is characteristic of each material and allows the unambiguous identification of the same. Thus, after obtaining Raman spectrum of a sample taken and the position of the bands appearing, compared with bands characteristics of pigments known patterns can be identified which one belongs. In the laboratory of TSC Raman spectroscopy has been obtained and stored the Raman spectra of a large collection of pigment pattern, which provides an extensive database to compare with the information contained in the Raman spectrum of a material under analysis . However, obtaining this information Raman is not always easy because the measured spectrum contains, in addition to the Raman signal, noise inherent to the process so far as to the nature of the material analyzed, which could mask the signal Raman. To improve the relationship signal a Raman spectrum can be used various techniques and signal processing, including in this thesis investigates the usefulness of processing techniques based on fuzzy logic. In this paper we have used this technique to filter processing, thereby improving the relationship signal in Raman spectra since it allows dealing with the uncertainty introduced by the noise inherent in the measure itself, and processing spectrum without loss of information, ie, without being modified the position of Raman bands. Moreover, the identification of a pictorial material is usually done by the analyst through visual inspection of the spectrum, the location of the bands in the spectrum and comparing these bands with the database features bands of pigment patterns that becomes available. This task is not easy and requires an experienced observer, in the light of their experience and knowledge gained by doing the work location and identification. In this process is introduced, inevitably, an element of subjectivity associated with the interaction of the investigator. Further work is often a costly and requires a high consumption of time. To try to avoid mistakes that can lead to work methodology, in this thesis has been used the potential of fuzzy logic to develop a system that is capable of carrying out the location automatic Raman bands and their subsequent identification by the c 8 omparaci 32b which is a database of patterns, thereby eliminating the element of subjectivity and therefore independent of the location and identification of trial investigator.
