Multi-Modal AI-Powered Crop Health Monitoring and Real-Time Diagnostic System

In modern precision agriculture, early plant pathology identification and rapid diagnostic decision-making areparamount to preventing catastrophic crop loss and securing food production. This project presents a Multi-Modal AI-Powered Crop Health Monitoring, Pathology Indexing, and Real-Time Diagnostic System designed for agricultural practitioners. The core engine integrates advanced computer vision using OpenCV for HSVcolor space segmentation alongside YOLOv8 deep learning instance segmentation to perform accurate leafnomaly extractions. By systematically quantifying infected leaf areas against total surface area, the system computes the Disease Severity Index (DSI %) and classifies pathogen strains, including Bacterial Leaf Blight,Powdery Mildew, and Yellow Leaf Spot, in real time. The application features a modular desktop interface built with Python and Tkinter, facilitating smooth image and continuous video stream processing with stabilized frame rates. To bridge technical diagnostic outputs with end-user accessibility, the system incorporates dynamic bilingual report generation in both English and Tamil. Furthermore, a thread-optimized Text-to-Speech (TTS) voice engine delivers real-time localized agronomic advice, contagion risk indicators, and precautionary recommendations without interface latency. Overall, this multi-modal diagnostic engineprovides a transparent, efficient, and scalable solution for precision crop health indexing and automated field diagnostics.