APPLICATION OF NIR SPECTROSCOPY AND DATA SCIENCE FOR NON-DESTRUCTIVE EVALUATION OF BARLEY (Hordeum vulgare L.)

• Grant number: 24/09561-0
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Effective date (Start): October 01, 2024
• Effective date (End): September 30, 2025
• Field of knowledge: Agronomical Sciences - Food Science and Technology - Food Science
• Principal Investigator: Juliana Azevedo Lima Pallone
• Grantee: Gabriel Gozzi
• Host Institution: Faculdade de Engenharia de Alimentos (FEA). Universidade Estadual de Campinas (UNICAMP). Campinas , SP, Brazil

Abstract

The research aims to explore the potential of using near-infrared (NIR) spectroscopy in portable and benchtop equipment combined with data science such as chemically interpretable machine learning algorithms, chemometrics, in non-destructive quality control of barley. Industry 4.0 and Green Chemistry are considered important guidelines in this context. Spectroanalytical techniques, such as NIR, in conjunction with multivariate machine learning algorithms (chemometrics) show promise for fast and accurate analyses. This approach can optimize laboratory processes, reduce costs and improve efficiency and product quality. However, there are few studies on the application of these techniques in barley control, focusing on comparisons between different algorithms and different spectrometer portabilities. For the study, samples of brewing barley and forage barley grains will be acquired to discriminate between both types of grains, authenticate against fraud due to mixtures and predict protein content. Samples will be obtained from different cultivars and regions to ensure data variability. The spectra will be obtained in the near-infrared region using benchtop equipment and microNIR equipment. The spectral data will be analyzed with linear and non-linear machine learning techniques for regression (PLS and SVM), discrimination (PLS-DA and SVM Classification) and authentication (DD-SIMCA and One-class SVM), using appropriate pre-processing according to the obtained spectral profile and data sets external to training to evaluate and validate the quality of the models. The results obtained can direct Industry 4.0 with the implementation of faster, more sustainable and efficient methods for barley quality, with an emphasis on important parameters for the brewing industry.