Cocoscan: AI-Powered Precision Diagnostics for Coconut Leaf Disease

Coconut cultivation is essential in tropical areas, but diseases such as Leaf Spot, Leaf Blight, Grey Leaf Spot, Root Wilt, Stem Bleeding, and Basal Stem Rot can severely harm coconut trees. To reduce the effects of these illnesses and guarantee the highest possible crop yields, early disease detection is important. Conventional detection techniques, which depend on specialist knowledge and manual examination, are labor-intensive, prone to human error, and frequently out of reach for smallholder farmers. This research presents a web-based application that automatically detects and categories diseases of coconut leaves using deep learning, most especially the Inception V3 model. The number of epochs to guarantee comprehensive training, a batch size that strikes a compromise between training speed and memory consumption, and a learning rate optimised for efficient convergence are important model parameters. In order to accommodate the architecture of the model, the image input size is set at 299×299 pixels. Through the web interface, farmers and agricultural specialists can submit photos of coconut leaves. The Inception V3 model, which has been trained on a sizable dataset of both healthy and diseased leaves, analyses the photos and offers real-time feedback, including suggestions for treatment, early disease detection, and disease identification. The user-friendly design requires no specialized knowledge, allowing farmers to quickly receive accurate diagnoses and preventive measures, improving farm management. The system also integrates Explainable AI (XAI) techniques to provide visual explanations of predictions, ensuring transparency and trust in the model's outputs.