A Decision Support System for Detection, Analysis and Classification of Paddy Plant Diseases

Paddy crop health is increasingly threatened by diseases that lead to significant yield loss, making manual inspection insufficient due to its time-consuming and unreliable nature. This paper proposes an end-to-end solution integrating an Android app and web interface for user interaction, a machine learning algorithm for disease classification, and cloud-based infrastructure for data storage and analysis. The system employs image processing and a Convolutional Neural Network (CNN) to analyze paddy leaf images, targeting four diseases: blast, bacterial leaf blight, brown spots, and tungro. Two models were developed—one to identify healthy crops and another to classify diseases. Farmers can capture and upload crop images via the Android app, which are then classified using a Google Cloud function. The system also stores telemetry data, such as location and images, for further analysis, including identifying disease hotspots and providing management recommendations. The app displays the nearest Krishi Vigyan Kendra (KVK) location and contact details using OpenStreetMap. With 95% accuracy in detecting healthy crops and 99% in disease classification, this solution aims to improve crop yield through early disease detection and timely interventions.