Detecting Alternaria solani in Tomatoes: Identification with VGG-19 Deep Learning for Early Detection

This paper addresses the crucial issue of early tomato disease detection, with a specific focus on Alternaria solani, by employing both machine learning and deep learning techniques. Our study extensively explores the capabilities of the VGG-19 convolutional neural network, a modified version of VGG-16, in the realm of image classification. The research methodology leverages a comprehensive dataset comprising over 2200 images sourced from the Plant Village dataset, with a primary emphasis on early blight disease. The core contribution of this work lies in evaluating the classification accuracy achieved by our proposed model. We report a remarkable accuracy of approximately 92% in distinguishing individual tomatoes as either healthy or affected by early blight disease. These results underscore the promising potential of our model in the identification of diseased tomatoes, which holds significant implications for agricultural disease management.