Evaluation and characterization of cocoa beans and cocoa honey (Theobroma cacao) by hyperspectral imaging, NIR spectroscopy and electronic nose combined with chemometrics

• Grant number: 24/22967-5
• Support Opportunities: Scholarships in Brazil - Doctorate
• Start date: October 01, 2025
• End date: July 31, 2028
• Field of knowledge: Agronomical Sciences - Food Science and Technology - Food Engineering
• Principal Investigator: Douglas Fernandes Barbin
• Grantee: Luis Alejandro Marzano Barreda
• Host Institution: Faculdade de Engenharia de Alimentos (FEA). Universidade Estadual de Campinas (UNICAMP). Campinas , SP, Brazil
• Associated research grant: 23/16840-0 - Integration of process analytical technologies to the cocoa processing chain, AP.R

Abstract

Reliable and non-destructive analytical methods for analyzing food and agricultural products are highly desirable, and near-infrared (NIR) spectroscopy, hyperspectral imaging and electronic nose stand out as fast and non-destructive alternatives. With advances and cost reduction in equipment and data science, new methods have been developed for food analysis, requiring research to evaluate their applications and suitability for different products. In this work, we propose to compare NIR spectroscopy, hyperspectral imaging and electronic nose to characterize the composition of cocoa pulp and honey, and fermented almonds. To analyze the data produced by these equipments, the use of chemometrics will be essential. Principal component analysis (PCA) will be used to evaluate the influence of fermentation on the spectra and volatile compounds of cocoa almonds. In addition to this method, the predictive method by partial least squares (PLS) will be used to obtain calibration and prediction models of the attributes measured experimentally through traditional analytical methods, and classification methods. The classification methods to be tested will be linear discriminant analysis (LDA), partial least squares linear discriminant analysis (PLS-DA) and soft independent modelling of class analogies (SIMCA). This project is expected to compare the performance of portable NIR spectrometers, hyperspectral imaging and electronic noses, combined with data science, to identify new methods for evaluating cocoa products.