A HYBRID IOT AND DEEP LEARNING FRAMEWORK FOR REAL TIME PLANT LEAF DISEASE DETECTION IN SMART AGRICULTURE

The rapid growth of the global population and increasing demand for food security have accelerated the adoption of smart agriculture technologies. Among the most critical challenges in modern farming is the early and accurate detection of plant diseases, which significantly reduce crop productivity and threaten sustainable agricultural practices. This research proposes a hybrid framework integrating Deep Learning (DL) and the Internet of Things (IoT) for real-time plant leaf disease detection. The system employs Convolutional Neural Networks (CNNs) fine-tuned through transfer learning to classify plant leaf images captured using IoT-enabled cameras and sensor networks. Lightweight CNN architectures such as MobileNetV2 are deployed on edge devices to enable fast, resource-efficient inference in field environments, while high-capacity models including ResNet50 and DenseNet121 operate in cloud servers for large-scale analysis and continuous retraining. Environmental sensor data (temperature, humidity, soil moisture) are also incorporated to contextualize disease prediction. Experimental evaluation was conducted using benchmark leaf disease datasets combined with IoT field-simulated inputs. Results demonstrate that DenseNet121 achieved the highest classification accuracy of 97.8%, while MobileNetV2 provided competitive performance (94.6%) with significantly reduced inference latency, making it suitable for edge deployment. The hybrid edge–cloud architecture ensures scalability, reduced response time, and improved accessibility for both smallholder and industrial farms. This work contributes to precision agriculture by addressing affordability, interpretability, and real-time scalability challenges in plant disease management.