This work discusses the application of the Near Infrared Vibrational Spectroscopy (NIR) technique in cocoa samples from different regions of the state of Pará (Medicilandia, Tucumă and Tomé-açu) with the aim of predicting major components and discriminating these different geographic regions. The NIR technique was associated with exploratory analysis of spectral data, testing different regression and discrimination models. This research is divided into three chapters. The first chapter reviews the concepts related to the NIR technique, and explores the most diverse articles on the application of the NIR technique in cocoa for different evaluations, such as, for analysis of major components, identification of fraud and discrimination of cocoa in different regions, from the planet. In the second chapter, prediction models were developed for nitrogen, moisture and total lipids in non-fermented samples, since certain cocoa components are not altered during fermentation, so these major components of the seeds could be quickly predicted by FT spectroscopy-NIR using PLS regression. The third article presents the application of NIR associated with exploratory analysis of spectral data by principal components, for discrimination of cocoa beans under five different treatment conditions (raw, dry fermented, dry unfermented and fermented defatted), from the three different regions mentioned geographic areas, and the best results were obtained when evaluating the fermented cocoa samples from the three regions.
|
