Tomato Leaf Disease Detection Based on Convolutional Neural Network

Abstract

AbstractTomatoes are a popular and important crop in India, with a large economic price and great production capacity. Diseases harm the health of the plant, which has an impact on its growth. It is critical to monitor the progress of the farmed crop to guarantee minimal losses. There are a slew of tomato diseases that are wreaking havoc on the crop’s leaves. One of the major linkages in the avoidance and control of crop diseases is the identification of infections in the leaf portions during the planting phase. Tomato leaves, including six popular species (Bacterial Spot, Black Mold, Early Blight, Late Blight, Mosaic Virus, and Septoria Spot), are used as experimental objects in this work to extract disease features from the leaf surface. Deep learning-based disease identification might help prevent such a catastrophe. A Convolutional Neural Network (CNN) is a type of deep learning algorithm that is currently commonly used for image categorization. In our studies, we used the CNN architecture to identify diseases in tomato leaves. This data set contains 2800 pictures of plant diseases. The Convolutional Neural Network was used in our proposed system to detect plant leaf diseases in seven categories, comprising six classes for diseases found in various plants and one class for healthy leaves. As a result, we were able to attain remarkable training and testing accuracy, with a training accuracy of 97.190% and a testing accuracy of 96.607% for all data sets used.KeywordsDeep learningLeaf diseasesConvolutional neural networkTransfer learning