Innovations in post-harvest disease detection: From molecular diagnostics to AI-based imaging

Abstract

Post-harvest diseases are a major contributor to global food losses, accounting for 20-50% of perishable crops, thereby threatening food security and economic stability. Traditional disease detection methods, such as visual inspection and microbiological culturing, are often slow, subjective, and lack the sensitivity needed for early pathogen identification. Recent advancements in biotechnology and computational analytics have introduced transformative solutions, including molecular diagnostics, spectroscopic techniques, and artificial intelligence-powered imaging systems. Molecular methods such as polymerase chain reaction, loop-mediated isothermal amplification, and CRISPR-based assays enable rapid and precise pathogen detection at the genetic level. Meanwhile, non-destructive technologies like near-infrared spectroscopy and hyperspectral imaging capture biochemical and morphological changes in produce, allowing for real-time monitoring. AI and machine learning further enhance these approaches by automating disease recognition through deep learning models such as convolutional neural networks, improving accuracy and scalability. This review comprehensively examines these innovations, discussing their principles, applications, advantages, and current limitations. Additionally, it explores future trends, including the integration of multi-modal detection systems and edge computing for on-site diagnostics. By leveraging these cutting-edge technologies, the agricultural sector can significantly reduce post-harvest losses, enhance food safety, and optimize supply chain efficiency.

Keywords: polymerase chain reaction, loop-mediated isothermal amplification, CRISPR, hyperspectral imaging, near-infrared spectroscopy, artificial intelligence, machine learning, deep learning, convolutional neural networks, food security, pathogen detection, non-destructive testing