Local Winter Vegetable Recognition using Deep Learning Techniques

Abstract

This study focuses on the use of deep learning techniques to recognize local winter vegetables, which is a vital part of agricultural precision farming. Tomato Leaf Blight Disease, Eggplant Healthy Leaf, Potato Leaf Anthracnose Disease, Potato Healthy Leaf, Bean Leaf Golden Mosaic Disease, Melon Leaf Late Blight Disease, Eggplant Leaf Blight Disease, Bitter Melon Healthy Leaf, Tomato Healthy Leaf, and Bean Healthy Leaf were among the ten target attributes carefully collected. A number of modern deep learning algorithms, including 'DenseNet201,' 'ResNet50,' 'VGG19,' 'InceptionV3,' and a standard Convolutional Neural Network (CNN), were used to achieve able recognition. Using the collected dataset, each algorithm received accurate training and evaluation processes. unexpectedly DenseNet201 developed as the most effective model, outperforming all others with a remarkable 99.68% accuracy. This success shows the importance of connected convolutional networks in capturing complicated patterns in a diverse and complex dataset of local winter vegetables. DenseNet201's high accuracy shows the capacity to recognize little changes in vegetable attributes, making it a powerful tool for agricultural applications. This study's findings not only contribute to the field of local winter vegetable recognition, but also indicate the efficacy of specific deep learning algorithms in addressing the complicated issues of plant health assessment. Because of DenseNet201's shown performance, there are positive opportunities for the creation of accurate and useful precision agriculture solutions, which will eventually help farmers make decisions more quickly and properly