Laser-based classification of olive oils assisted by machine learning

Area of Science:

• Analytical Chemistry
• Spectroscopy
• Machine Learning

Background:

• Olive oil is vital to Mediterranean economies, producing 90% of global supply.
• Quality assessment and origin authentication (PDO/PGI) are crucial for the olive oil market.
• Current methods for olive oil analysis can be time-consuming.

Purpose of the Study:

• To develop a rapid, reliable method for olive oil classification.
• To assess olive oil acidity and geographical origin using a novel technique.
• To enhance quality control and authentication processes in the olive oil industry.

Main Methods:

• Utilized Laser Induced Breakdown Spectroscopy (LIBS) for elemental analysis of olive oil.
• Applied machine learning algorithms to interpret LIBS spectral data.
• Developed a methodology for classifying olive oil based on acidity and origin.

Main Results:

• Demonstrated LIBS combined with machine learning as an ultrafast classification tool.
• Successfully differentiated olive oils based on their degree of acidity.
• Enabled authentication of olive oils by geographical origin (PDO/PGI).

Conclusions:

• LIBS and machine learning provide a powerful, fast, and in-situ solution for olive oil quality control.
• This approach supports accurate olive oil authentication and market surveillance.
• The methodology offers a significant advancement for the global olive oil industry.