A group of researchers from the University of Bari and Milan has developed and tested a non-destructive analysis system based on chemometric and artificial intelligence approaches applied to infrared spectroscopy, to evaluate the quality extra virgin olive oil quickly, sustainably and at low cost.
The study, published in the journal Food Chemistry, proposes a protocol capable of estimate the concentration of fatty acid ethyl esters, key indicators for the quality and authenticity of extra virgin olive oil.
Currently, this parameter is determined using gas chromatography, a reliable but complex, slow, and expensive procedure that requires the use of chemical reagents and equipped laboratories.
The new method uses and valorises, instead, the information obtained from FT-IR spectroscopy, a sort of fingerprint of the product, through multivariate analysis and machine learning models capable of identifying correlations invisible to the human eye.
THEmore effective algorithm, based on the XGBoost technique, thanks also to the use of explainable artificial intelligence tools, allows to identify and interpret the spectral regions most associated with the presence of ethyl esters.
The findings thus contribute to the paradigm shift affecting quality control in the food supply chain, and the olive oil and olive oil sector in particular, which is still one of the most affected by adulteration and fraud to the detriment of consumers.
This technology, in fact, It will allow to drastically reduce the time and costs of analysis, reduce the environmental impact linked to traditional methods and carry out rapid screening on larger quantities of samples, providing an immediate, reliable indication of product compliance.
Although the approach does not yet replace the official method, it can immediately become an effective preliminary tool for producers, mills, consortia, and certification bodies. “Our ambition” explains the scientific director of the METRDOFOOD-IT project for the University of Bari, Professor Sabina Tangaro, “is to provide the olive oil sector with an intelligent, fast and sustainable tool, capable of concretely improving quality control processes. The combination of spectroscopy and artificial intelligence has the potential to revolutionize this field, making testing more accessible and efficient.”
The research team is already working on expanding the experimental dataset and extending the methodology to other extra virgin olive oil quality parameters, such as acidity, peroxide value, and phenolic content, with the aim of developing an integrated system capable of offering a comprehensive product assessment.
The result is also part of the METROFOOD-IT project, in which the University of Bari plays a strategic role in the study and implementation of artificial intelligence models applied to the agri-food sector, contributing to the development of advanced tools for the evaluation and quality control of food production.
