Detection of Meat Adulteration Using Spectroscopy-Based Sensors

Area of Science:

• Food Science
• Analytical Chemistry
• Spectroscopy

Background:

• Minced meat is susceptible to economically motivated adulteration due to the difficulty in identifying species and tissue origins.
• Morphological characteristics are often obscured in minced products, making visual inspection unreliable.

Purpose of the Study:

• To evaluate the effectiveness of spectroscopy-based sensors in detecting adulteration in minced meat.
• Specifically investigating the substitution of beef with bovine offal and pork with chicken (and vice versa).

Main Methods:

• Preparation of minced meat samples with varying adulteration levels (25% increments) of beef/bovine offal and pork/chicken mixtures.
• Acquisition of visible (Vis), fluorescence (Fluo) spectra, and multispectral images (MSI) from 120 fresh and frozen-thawed samples.
• Development and evaluation of Support Vector Machine (SVM) classification models for each sensor type.

Main Results:

• Multispectral imaging (MSI) models achieved the highest accuracy, ranging from 87% to 100%.
• Visible (Vis) spectroscopy models demonstrated good accuracy (57% to 97%), outperforming fluorescence (Fluo) models.
• Sensor-specific performance was observed, with MSI being the most effective detection method.

Conclusions:

• Spectroscopic data, particularly MSI, show significant potential for detecting and quantifying minced meat adulteration.
• The choice of sensor is crucial for effective fraud detection in food commodities like minced meat.