Durain Leaf Disease Identification using Deep Learning

Project Code :TCMAPY2277

Objective

The Durian Leaf Disease Identification project uses deep learning techniques, specifically MobileNet and EfficientNet, to accurately detect and diagnose leaf diseases in durian crops. By leveraging these algorithms, the system classifies leaf images and provides insights into disease severity. The user-friendly web interface, built with Flask and MySQL, allows users to upload images for analysis. The tool aids farmers and agricultural experts in making informed decisions about crop treatment, improving disease detection speed and agricultural productivity for healthier crops and better yields.

Abstract

This project develops an automated web-based system for detecting and classifying diseases in durian leaves using deep learning. It identifies 10 common durian conditions — anthracnose disease, canker disease, fruit rot, mealybug infestation, pink disease, sooty mold, stem blight, stem cracking with gummosis, thrips disease, and yellow leaf — through two efficient PyTorch models: MobileNetV2 (lightweight and fast) and EfficientNet-B0 (higher accuracy with balanced computation). Users can upload leaf images via a Flask web application that includes registration, login, image upload, model selection (MobileNet, EfficientNet, or both), and real-time prediction showing the disease name, confidence score, and top-3 likely classes. Images are preprocessed to 224×224 resolution with ImageNet normalization before inference. The system aims to provide farmers and agronomists with quick, reliable, accessible disease diagnosis support for better durian crop management and yield protection.

Keywords: durian leaf disease, disease classification, deep learning, CNN, EfficientNet, MobileNet, PyTorch, Flask, web application, precision agriculture

NOTE: Without the concern of our team, please don't submit to the college. This Abstract varies based on student requirements.