Artificial Intelligence–Assisted Near Infrared Spectroscopy for Dynamic Process Monitoring and Control in the Food Industry: Current Advances, Challenges, and Future Perspectives
Food and Bioprocess Technology, cilt.19, sa.9, 2026 (SCI-Expanded, Scopus)
- Yayın Türü: Makale / Derleme
- Cilt numarası: 19 Sayı: 9
- Basım Tarihi: 2026
- Doi Numarası: 10.1007/s11947-026-04530-8
- Dergi Adı: Food and Bioprocess Technology
- Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Compendex, Food Science & Technology Abstracts, INSPEC, Academic Search Ultimate (EBSCO), Natural Science Collection (ProQuest), Engineering Source (EBSCO), Materials Science & Engineering Collection (ProQuest), Technology Collection (ProQuest)
- Anahtar Kelimeler: Artificial intelligence, Digitalization, Food process monitoring, Industry 4.0, On-line and in-line near infrared spectroscopy, Process analytical technology
- Hacettepe Üniversitesi Adresli: Evet
Özet
The growing global demand for safe, sustainable, and high-quality food products has increased the need for real-time, non-destructive process monitoring technologies capable of supporting intelligent food manufacturing. Near-infrared spectroscopy (NIRS) has emerged as one of the most versatile process analytical technologies (PAT) for the food industry, offering rapid, continuous, non-destructive, and multi-parameter measurements that are well suited for dynamic process control. When integrated with Industry 4.0 technologies—including Internet of Things (IoT)–enabled sensor networks, cyber-physical systems, and artificial intelligence (AI)—NIRS supports predictive process monitoring and adaptive process optimization. This review critically evaluates recent developments in the field of AI-enabled NIRS for dynamic process control across the food supply chain; this evaluation places particular emphasis on measurement configurations (on-line and in-line), data analysis strategies, industrial applications, current limitations, and future research directions. Furthermore, it critically evaluates the current limitations and knowledge gaps that hinder the industrial applications of AI-enabled NIRS and identifies future research priorities aimed at improving model robustness, transferability, and process reliability. The reviewed studies suggest that AI-based models can enhance the predictive capability of NIRS for complex food systems. Nevertheless, robust industrial implementation remains challenged by calibration robustness, model transferability, sensor variability, and the lack of standardized validation frameworks. Emerging developments, including transfer learning, multi-sensor data fusion, digital twins, federated learning, and foundation-model-assisted analytics, may facilitate the next generation of intelligent food manufacturing, although most remain at an early stage of industrial implementation. Overall, this review critically compares conventional chemometric and AI-based approaches for NIRS, highlighting current challenges and future research priorities for food process monitoring.