Internet of Things and Human Neocortex Inspired Algorithms for the Plant Disease Prediction-Sheath Blight for Paddy Plant

Abstract

Sheath blight disease (Rhizoctonia solani) poses a critical threat to paddy (Oryza sativa) production, causing up to 50% yield losses under conducive environmental conditions. Existing computer vision-based disease detection systems identify infections only after symptom manifestation, while conventional machine learning prediction models suffer from high computational demands, data preprocessing requirements, and noise sensitivity. This research proposes a novel Hierarchical Temporal Memory (HTM) framework for real-time, pre-symptomatic disease prediction using Internet of Things (IoT)-based environmental monitoring of agricultural fields. Inspired by human neocortex architecture, HTM exhibits inherent noise resistance and continuous learning capabilities without extensive retraining. The proposed system leverages three critical environmental parameters, temperature, humidity, and rainfall, collected via IoT sensors to predict disease onset before visible symptom development. Implementation and validation were conducted from 2019 to 2023 for sheath blight prediction in paddy cultivation, achieving 94% prediction accuracy in 2023. This pre-emptive prediction capability enables timely intervention, reduced pesticide application, and enhanced sustainable agricultural practices. The HTM-IoT integration represents a significant advancement in precision agriculture by transitioning from reactive disease detection to proactive disease forecasting, supporting both crop productivity and environmental sustainability objectives.