Crop Disease Identifier from Leaf Images

Abstract

The Crop Disease Identifier from Leaf Images is a deep  learning-based system designed to 
detect crop diseases accurately and efficiently. It utilizes a ResNet9 CNN model implemented 
in PyTorch, trained on a dataset of 38 healthy and diseased leaf classes across major crops 
like apple, corn, grape, potato, and tomato. Image preprocessing techniques such as resizing, 
normalization, and augmentation improve model generalization. The trained model is 
integrated into a Flask web application, allowing users to upload leaf images for real-time 
disease prediction and treatment suggestions. The system achieves high accuracy, supports 
scalability for additional crops, and runs efficiently on standard and mobile devices. This 
project provides a cost-effective, farmer-friendly, and scalable solution to enhance sustainable 
agriculture by minimizing losses and reducing excessive pesticide use.