An Efficient Model for Plant Disease Detection in Agriculture Using Deep Learning Approaches

Abstract

Detecting plant diseases is crucial as they can significantly impact plant growth. At the same time, several machine-learning methodologies have been employed to distinguish and categorize plant diseases. Deep learning, a subset of machine learning, has shown promising results in terms of accuracy. Using CNN architectures and visualization methods, scientists can develop and implement effective strategies for detecting and classifying plant disease symptoms. These methods primarily rely on analyzing leaf characteristics such as color and damage. A plant-village dataset of 54,306 images representing 14 different crop classes and 26 types of infections was used to train the models. The convolutional neural networks (CNNs) and additional machine learning models were trained on this dataset and achieved an impressive accuracy rate of 95% in identifying disease symptoms on plant leaves. Additionally, an efficient web-based application for crop disease detection was developed as part of this approach.