Plant Leaf Disease Prediction and Classification Using Deep Learning
Meroua Belmir, Wafa Difallah, Abdelkader Ghazli
Abstract
One of the primary causes of decreased agricultural production in terms of both quality and quantity is the prevalence of plant diseases. Plant diseases continuously emerge on plant leaves due to ongoing changes in plant structure and cultivation methods. Initially, these diseases affect the leaves and subsequently spread throughout the entire plant, resulting in a significant impact on the variety and yield of crops that can be grown. Therefore, effective control of infection spread and promotion of robust plant growth relies on accurate classification and early detection of plant leaf diseases. In this study, a deep convolutional neural network (CNN) model is proposed for the classification of plant diseases, aiming to address these challenges. The model utilizes the PlantVillage dataset, which consists of images showcasing 14 different healthy and diseased crop leaves, categorized into 38 distinct classes. Experimental results demonstrate that the CNN model achieves a training accuracy of 98.01 % and a test accuracy of 94.33%. These findings highlight that the model offers a highly effective solution for the early diagnosis of leaf diseases.