Technological Innovations to Reduce Post‑Harvest Grain Losses

Abstract

Food security is a critical global challenge and post-harvest grain losses present a significant threat, with 9-18% of the world's cereal grain production lost from harvest to consumption. We have explored and validated a novel monitoring, prediction and traceability system of Internet of Things (IoT) sensors, artificial intelligence (AI) and blockchain distributed ledger. IoT sensors continuously monitored the temperature, relative humidity, carbon dioxide and acoustic insect numbers in grain silos for 120 Nine machine learning algorithms were trained to predict dry matter loss (DML) and early spoilage. Our Random Forest model (RMSE = 0.043%, R² = 0.967) was the best performing with Long Short Sensor fusion of the four sensors types reduced the ξ The δ.