Disorder Detection in Tomato Plant Using Deep Learning

Abstract

Agricultural productivity is something on which the Economy highly depends. Plant Diseases and Pests are a major provocation in the agriculture sector. This is one of the reasons that disease detection in plants plays an important role. Accurate and faster detection of diseases in plants could help to develop an early treatment technique while substantially reducing economic losses. This paper proposes a deep learning-based approach that automates the process of classifying tomato leaves diseases. The proposed system focuses on major diseases like early blight, powdery mildew and downy mildew that occur in tomato plants. We make use of Convolution Neural Network to classify the image data sets based on the visible effects of diseases. Unlike Image Processing, Deep Learning learns and adapts to the changing data sets. The proposed methodology will be using a diverse dataset that includes images from the nursery, plant village, and farm. To test the diverse dataset, we use Receiver Operating Characteristic that will improve the classifier performance. Thus, the proposed system serves as a phone based tool that helps in detecting tomato plant diseases based on capturing and analyzing the picture of a plant leaf. In this paper, we take a first step towards such a tool. As a result, the proposed system serves as a decision support tool to farmers in identifying diseases that occur in tomato plant leaf.