Detection of fat and chicken adulteration in lamb using VIS/NIR spectroscopy and LDA model

Document Type : Research Paper

Authors

• Amit Kazemi ¹
• Asghar Mahmoudi ²
• Hadi Veladi ³
• Arash Javanmard ⁴

¹ Department of Biosystem engineering, faculty of agriculture, Tabriz university, Tabriz, Iran

² Biosystem engineering department , agriculture faculty . university of Tabriz , Iran

³ Department of Electrical and computer engineering, university of Tabriz, Tabriz, Iran

⁴ Department of Animal Science, Faculty of Agriculture, University of Tabriz,

Abstract

Keywords

• Spectroscopy
• adulteration
• meat
• PCA
• LDA

Main Subjects

• food industry

EXTENDED ABSTRACT

Introduction

Meat is one of the essential food products for the human diet, which is widely consumed by consumers due to its nutritional value and palatability. According to the Food and Agriculture Organization (FAO), there has been a growing demand for high-value meats such as beef and lamb. To meet the increasing consumer expectations for the production of safe, high-quality, and cost-effective products, scientists are exploring some methods to ensure meat quality.

Authenticity is a crucial quality factor for desirability. Several traditional methods exist for assessing meat safety and quality.

For this reason, non-destructive and rapid methods are being investigated to improve efficiency in detecting fraud in the meat industry. Vibrational spectroscopy is one of the most common analytical methods for assessing the quality of meat and meat products in recent years. Infrared spectroscopy is a vibrational spectroscopy technique based on the interaction between infrared waves and the material under study. In this study, the extent of adulteration in minced lamb meat was determined using NIR spectroscopy and chemometric methods (statistical chemistry). The adulteration involved mixing lamb fat or chicken meat with lamb meat. After conducting experiments and relevant analyses, the efficiency of the method was evaluated.

Material and methods

This study investigates the detection of adulteration in lamb meat with fat and chicken meat using VIS/NIR spectroscopy in the wavelength range of 200 to 1100 nm. Adulterated samples were manually prepared with weight-based adulteration levels of 5%, 10%, 15%, and 20%. A total of 190 samples were examined, each weighing 10 grams. To eliminate additive and scattering effects in the spectral data, various preprocessing methods, including derivatives and scattering correction techniques, were applied. Principal Component Analysis (PCA) was used for variable reduction and data clustering, while Linear Discriminant Analysis (LDA) models, combined with different preprocessing approaches, were employed to classify the meat samples.

Results

The best prediction accuracy was achieved using Savitzky-Golay preprocessing with the accuracy of 70.76% for the 9-class data and 92.89% for the 3-class datasets. The best clustering results were obtained using the PCA model with Savitzky-Golay (SG) preprocessing. The first two principal components (PCs) captured the majority of the variance within the dataset (95%).

Discussion

One of the key findings of this study is that spectroscopy, as a non-destructive method, is capable of detecting low levels of adulteration without the need for sample destruction. However, it is recommended that future research focus on enhancing the accuracy and practical applicability of this technique in industrial or regulatory settings by expanding sample diversity, employing deep learning models, and investigating the most informative spectral regions for sample discrimination.

Author Contributions

Amir Kazemi: Conceptualization, Methodology, Software, Formal analysis, Investigation, Resources, Data curation, Writing–review & editing, Project administration, Writing-Original Draf.

Asghar Mahmoudi: Project administration, Investigation, software, supervision, validation,

Hadi Veladi: Methodology, Investigation, Data curation, Investigation

Arash Javanmard: Methodology, Software, Writing–review & editing

All authors have read and agreed to the published version of the manuscript.

Data Availability

Data available on request from the authors.

Acknowledgements

The authors gratefully acknowledge Dr. Mostafa Khojastehnazhand for his instructive suggestions during preparation of first draft of manuscript.

Ethical considerations

The authors avoided data fabrication, falsification, plagiarism, and misconduct.

Conflict of interest

The author declares no conflict of interest.