Citrus Leaf Disease Classification Using VGG16 Convolutional Neural Network Architecture

Abstract

This study focuses on the importance of the classification of citrus leaf diseases from images by utilizing the VGG16 Convolutional Neural Network (CNN) architecture. The objective of this research is to develop a strong framework for categorizing images of citrus leaf diseases, utilizing the demonstrated effectiveness of the VGG16 model in recent researches. The process entails categorizing a dataset of 607 photos into five unique classes, utilizing a training program consisting of 100 epochs, a batch size of 128, and a test size of 0.1. The investigation's results demonstrate the excellent performance of the proposed VGG16 CNN model, obtaining a noteworthy accuracy rate of 93.66%. This discovery highlights the model's expertise in acquiring complex patterns and essential characteristics necessary for differentiating between different types of citrus leaf diseases. This research has major significance, as it provides a dependable and automated tool for managing citrus trees. The achieved high accuracy indicates the possibility for practical application in precision agriculture, enabling early disease identification and supporting sustainable citrus production methods. This study contributes to the progress of technology-driven solutions for agricultural difficulties. The proposed CNN model is a great tool for growers rapid and precise disease surveillance in citrus plantations.