CLASSIFICATION OF TOMATO DISEASES USING ENSEMBLE LEARNING

Abstract

A Plant disease is any dysfunction of a plant, caused by living organisms, which affects the quality and quantity of yield. These symptoms are visually shown on the plant leaves. This paper discusses classification of Tomato diseases such as Late Blight, Septoria Leaf Spot and Yellow leaf curl virus while distinguishing the healthy leaf at the same time. An experimental sample size of 1817 was considered in conducting this study. This work differentiates diseased tomato leaf images with healthy leaf images. The classifiers Random Forest, Multilayer Perceptron Neural Network and Support Vector Machines were trained and got a prediction accuracy of 88.74%, 89.84%, and 92.86% respectively in classifying diseases. Then, the prediction results of Random Forest, Multilayer Perceptron and Support Vector Machines were combined using Soft Voting classifier and obtained a highest accuracy of 93.13% in classifying tomato diseases.

Keywords

• tomato diseases
• support vector machines
• multilayer perceptron
• random forest