Terahertz spectroscopy (0.1~10 THz), as a new type of non-destructive testing method with both microwave and infrared characteristics, has shown remarkable potential in the field of food quality testing in recent years. Its unique penetration, high sensitivity, and low photon energy characteristics, combined with chemometrics and machine learning methods, provide an efficient solution for the qualitative and quantitative analysis of complex food ingredients. In this paper, we systematically review the principles of terahertz spectroscopy and its key applications in food testing, focusing on its research progress in pesticide residues, additives, biotoxins, and mold, adulteration identification, variety identification, and nutrient content detection. By integrating spectral data preprocessing, reconstruction algorithms, and machine learning model optimization strategies, this paper further analyzes the advantages and challenges of this technology in enhancing detection accuracy and efficiency. In addition, combined with the urgent demand for fast and nondestructive technology in the field of food detection, the future development direction of the deep integration of terahertz spectroscopy technology and artificial intelligence is envisioned, with a view to providing theoretical support and technical reference for food safety assurance and nutritional health research.
Keywords: chemometrics; food safety and quality; machine learning; non-destructive testing; terahertz spectroscopy.