Tomato Leaf Diseases Detection and Classification using Convolutional Neural Network (CNN)

Abstract

Tomato (Lycopersiconesculentum Mill.) is one of the most grown vegetable plants in the world, second to potato. It originally came from the tropical area from Mexico to Peru. Tomato is grown in many parts of the country and also among the most important vegetable crops and its production has shown a marked increase. It became the most profitable crop providing a higher income to small scale farmers compared to other vegetable crops. There is a number of limits tomato yield some of those are caused by bacteria, vires, and fungus. Tomato leaf diseases are used to detect and classify symptoms of plant diseases, detection of tomato leaf diseases through the naked eye is ineffective, especially because there are numerous diseases. Therefore, we need to develop automatic tomato leaf diseases detection and classification. In this study we designed TomdiseasesNet model architecture based on inseption v3, VGG16, and mobile net architecture to perform tomato leaf diseases detection and classification using RGB, Grayscale, and augmented input image dataset. Deep learning become the most accurate and precise paradigms for the detection of plant diseases. Leaves of infected crops are collected and labeled according to the disease. Processing of image is performed along with pixel-wise operations to enhance the image. It is followed with feature extraction the classification of patterns of captured leaves in order to identify tomato plant leaf diseases. Six classifier labels are used as Tomato Bacterial Spot, Tomato Early Blight, Tomato Sectorial Leaf spot, Tomato Leaf mold, Tomato Yellow Leaf, and Healthy leaf. The features extracted are fit into the neural network with 200 epochs, 70/30 splitting ratio, and 0.0001 learning rate. TomdiseasesNet model achieved a testing accuracy 99.68%, Training accuracy 99.97%, and validation accuracy 99.78. The overall accuracy obtained 99.68% hence tomato plant leaf diseases infected by Tomato Bacterial Spot, Tomato Early Blight, Tomato Sectorial Leaf spot, Tomato Leaf mold, and Tomato Yellow Leaf can be detect the tomato plant leaf diseases and classify the types of diseases with a high rate of accuracy.