A rapid wine brand identification method based on the joint application of SERS and machine learning techniques

Area of Science:

• Analytical Chemistry
• Spectroscopy
• Machine Learning

Background:

• Accurate wine brand identification is crucial for authenticity and quality control.
• Traditional methods can be time-consuming or destructive.
• Surface-Enhanced Raman Scattering (SERS) offers high sensitivity for molecular detection.

Purpose of the Study:

• To develop a rapid, non-destructive method for red wine brand identification.
• To combine SERS spectroscopy with machine learning for enhanced accuracy.
• To enable real-time analysis for practical applications.

Main Methods:

• Utilized Surface-Enhanced Raman Scattering (SERS) spectroscopy to capture molecular signatures of red wine.
• Applied Principal Component Analysis (PCA) for spectral preprocessing.
• Employed a Convolutional Neural Network (CNN) for feature extraction and classification.

Main Results:

• Achieved a maximum classification accuracy of 99.27% across 18 red wine brands.
• Demonstrated real-time identification in under 3 seconds per sample.
• Validated the method on 1080 spectra from wines of diverse origins (China, Chile, Italy).

Conclusions:

• The SERS-CNN approach provides a highly accurate and efficient solution for red wine brand identification.
• This non-destructive technique is portable and suitable for various applications.
• Significant potential for wine brand protection, quality monitoring, and consumer markets.