Classification of Tomato Disease using Convolutional Neural Network

Abstract

Agriculture is an important sector for national food needs. The main crops in agriculture include rice, wheat, potatoes, sugar cane. Other crops include nuts, fruit, vegetables, tubers and others. Plant diseases are one of the problems that always arise in the agricultural sector. Every food ingredient has the potential to have disease. Usually diseased plants will be given pesticides. Detection of plant diseases is very important. Especially if we know the type of disease from the plant. To find out the type of disease with the naked eye is quite difficult, especially since the form of the disease has a similar pattern. In this research, we propose to classify the tomato disease from its leaf with thousands of tomato images. There are three types of analysed diseases they are bacterial spot, early blight and yellow leaf curl. We implement convolutional neural network approach to find the best classifier model. From various experiments, it was found that the neural network architecture that was built could achieve accuracy up to 87%