Converging Eco-Compatible Interventions with Predictive Artificial Intelligence (AI) in Post-Harvest Pulse Preservation: A Comprehensive Overview

Post-harvest deterioration substantially reduces the quality, safety, nutritional value, and marketability of stored pulses, particularly under fluctuating temperature and moisture conditions and increasing restrictions on conventional fumigants. However, the literature remains fragmented across pulse-deterioration mechanisms, preservation technologies, and emerging data-driven storage systems, with limited integration of pulse-specific evidence across these domains. This review addresses this gap by critically synthesizing and comparing conventional and emerging approaches, including hermetic and controlled-atmosphere storage, biological and botanical interventions, nano-enabled preservation, active packaging, irradiation, Internet of Things (IoT)-based monitoring, artificial intelligence (AI)/machine learning (ML), computer vision, and Digital Twin technologies. A structured literature synthesis indicates that hermetic and controlled-atmosphere storage are supported by documented pulse-specific preservation outcomes and established applications, whereas botanical, nano-enabled, and other emerging interventions show more variable validation across commodities and storage conditions. IoT and AI/ML show potential for environmental forecasting, quality assessment, and decision support, but pulse-specific longitudinal datasets, external validation, interoperability, and commercial-scale deployment remain limited. Overall, the evidence indicates that reliable next-generation pulse preservation will depend on integrating validated physical and biological interventions with progressively validated digital intelligence, supported by standardized datasets, multi-location and commercial-scale validation, techno-economic assessment, and appropriate safety and regulatory evaluation.