Abstract
This review article provides a comprehensive analysis of advanced technologies for detecting, analyzing, and controlling food spoilage, with a focus on perishable foods such as fruits, vegetables, and meats. Although traditional methods such as microbiological testing and sensory evaluation remain fundamental, emerging technologies such as machine learning (ML), computer vision, and electronic noses (enoses) offer transformative potential for real-time monitoring and predictive analytics. However, practical implementation of these technologies faces significant challenges, including heterogeneity in data, computational constraints, and environmental variability. For example, ML models, particularly deep learning architectures, require extensive labeled datasets and high-performance computing resources, which are often inaccessible in resource-constrained settings. Similarly, electronic noses, while effective in detecting volatile organic compounds (VOCs) associated with spoilage, suffer from sensor drift and cross-sensitivity issues, necessitating frequent recalibration. Blockchain technology, though promising for improving traceability and transparency in the food supply chain, struggles with scalability and energy efficiency. This review critically evaluates these limitations, highlighting gaps in current methodologies, such as the overreliance on external spoilage indicators in computer vision systems and the lack of standardized protocols for data collection and model evaluation. By addressing these challenges, future research can advance the development of robust, scalable and cost-effective solutions for food spoilage detection, ultimately contributing to improved food safety, reduced waste, and enhanced supply chain efficiency.