Using Machine Learning Techniques for Rice Leaf Disease Detection

Abstract

In order to prevent global warming on Earth, one of the most important resources is plants. Nonetheless, a multitude of ailments plague the plants. Just recently has study begun on the identification of plant diseases. To identify rice sickness is the major purpose of this article. Brown Spots Diseases, Leaf Blast Disease, and Leaf Blight Disease are a few illnesses that can affect rice. at different stages of growth. These infections impede the rice's ability to spread and protect its whole plant. Three different disease kinds were examined in this study along with one set of healthful rice leaves. Many different species, including fungi, bacteria, and other microorganisms, can cause rice disease. By cutting down on the amount of time required to ascertain the effect of rice leaf illness on humans, the technique was designed to consequently eliminate noise and produce the best results for leaf disease identification using ML with a greatest level of accuracy. This was achieved by applying ML techniques, including a computer-based detection approach. K-Fold validation procedures were used to measure the classification of this study. Random Forests, Decision Trees, Logistic Regression, and other support vector classifiers (SVCs) were trained with four classes of rice leaves. When K-fold cross validation tactics were applied to forecast 3 types rice leaf disease using one class of normal rice leaf, Random Forest produced the highest accuracy of 94.16%. At last, rice leaf detection through classification is accomplished using the CNN InceptionV3 model.