IoT and Inception V3 for Precision Rice Disease Detection

This paper shows how to use artificial intelligence and the Internet of Things (IoT) together to make an integrated smart farming framework that helps with precision agriculture in rice farming. The suggested system has a rice disease detection module that uses deep learning and the Inception V3 convolutional neural network (CNN). It also has a proximity-based irrigation control system that uses real-time soil data to turn on. We first trained the disease detection model with 3,000 real-field rice leaf images. Then, this method improved it with images from public datasets like PlantVillage, Mendeley, and Kaggle, which made it better at generalising across different conditions. The model had an F1 score of 98.06%, an accuracy of 98.12%, a precision of 98.93%, and a recall of 98.35%. Sensor nodes are used to continuously check the temperature, humidity, and moisture levels in the soil. These outputs are then combined into a low-cost IoT subsystem. An ultrasonic proximity sensor controls the irrigation system, which turns on when the soil moisture drops below 35%. This makes sure that the water is used efficiently. When used in the field, it showed a 22.4% increase in irrigation efficiency and an 18.7% increase in crop yield compared to traditional methods. The system is made to work well on small computers like the Raspberry Pi, making it a scalable and cost-effective solution for real-world farming. This study creates a new closed-loop crop monitoring system that connects AI diagnostics with sensor-driven actuation to make farming more sustainable.