Desenvolvimento de metodologias analíticas não destrutivas empregando imagens hiperespectrais no infravermelho próximo e Raman para análise de cruzamentos de traços de canetas

The determination of the chronological sequence of intersecting lines is an old problem in Questioned Documents forensic field that lasts until today. In Questioned Documents examinations, objective, fast and, mainly, non-destructive methodologies are preferable due to the need to keep the documents intact. In this work, near infrared (NIR) and Raman hyperspectral images associated with chemometrics techniques were used to determine the chronological sequence of intersecting lines made with blue and black pens. In NIR hyperspectral images analysis, different types of preprocessing techniques were tested and the images were treated using principal component analysis (PCA) and augmented multivariate curve resolution - alternating least squares (MCR-ALS). In most cases, it was not possible to differentiate ink from paper and just in 1 of the 15 intersection pairs analyzed it was possible to determine the correct chronological sequence of intersecting lines In Raman hyperspectral images analysis, spectra were pre-processed with asymmetric least squares (AsLS) and unit vector normalization and treated with k-means, MCR-ALS and partial least squares – discriminant analysis (PLS-DA). The results of these analysis were, first, analyzed visually, considering the quantity of ink and continuity of the lines. After, MCR-ALS and PLS-DA results were analyzed using a methodology based on the relative quantity of pixels of each ink in the intersection region, in order to make the determination objective. The obtained results were, then, compared with the results of optical microscopy analysis, commonly used technique, and it was verified that the objective methodology resulted in higher correct determinations rates. A 73% of correct determinations rate was obtained using MCR-ALS and the objective methodology. These results show that the proposed methodology is feasible to be adopted by the criminal investigators in the analysis of intersecting lines.