IRJET - Automatic Methods for Classi?cation of Plant Diseases using Back-Propagation Neural Networks

This document presents an automatic method for classifying plant diseases using back-propagation neural networks. The proposed method uses image processing techniques like preprocessing, feature extraction, and classification. Color and texture features are extracted from input images and used to train a back-propagation neural network for classification. The system is tested on 50 images of leaves with four types of disease symptoms and aims to automatically classify new test images. MATLAB is used to implement the preprocessing, feature extraction from gray level co-occurrence matrices, network training, and classification stages of the proposed method.