Abstract
Plant diseases pose a critical threat to global food security, agricultural sustainability, and farmer livelihoods, particularly in regions with limited access to advanced diagnostic technologies. Traditional methods of disease detection rely heavily on manual inspection, which is time-consuming, error-prone, and often results in delayed interventions. This paper presents a novel, solar-powered autonomous robotic system designed to detect plant diseases in real time using deep learning and IoT technologies. The proposed system integrates a high-resolution imaging unit, IoT-based environmental sensors, and an onboard processing module based on Raspberry Pi. Deep CNNs, trained on diverse datasets including PlantVillage, are used for accurate disease classification, while soil moisture and temperature sensors provide contextual environmental data to support diagnosis. The robot’s mobility, powered by solar energy, allows for continuous field monitoring with minimal human intervention. Experimental results demonstrate the system’s high classification performance, achieving 99.39% training accuracy, 97.47% validation accuracy, and 97.13% testing accuracy. Furthermore, the model achieved 99.63% overall accuracy, with a Precision of 99.40%, a Recall/Sensitivity of 99.56%, an F1-score of 99.46%, and a Specificity of 99.99% across multiple disease classes. These results highlight the robustness of the proposed approach in real-world agricultural conditions, enabling reliable disease detection and monitoring. The integration of cloud-based monitoring enables farmers to receive real-time alerts and insights, supporting timely and informed decision-making. This cost-effective, scalable, and environmentally sustainable solution has the potential to transform precision agriculture by enhancing early disease detection, reducing pesticide overuse, and improving crop yield and health.