Automatic classification of paddy leaf disease

Abstract

Rice is a staple food in most of the Asian countries. It is an important crop, and over half of the world population relies on it for food. However, paddy leaf disease can affect both the quality and quantity of paddy in agriculture production. The classification of paddy leaf disease is an important and urgent task as it destroys about 10% to 15% of production in Asia. Thus, a study on automatic classification of paddy leaf disease using image processing is presented. Feature extraction techniques of color, texture, and shape were implemented to analyze the characteristics of the paddy leaf disease. In another note, a Support Vector Machine (SVM) is used to classify the four types of paddy leaf disease which are the brown spot, bacterial leaf blight, tungro virus, and leaf scald. The performance of the proposed study is evaluated to 160 testing images which returned 86.25% of classification accuracy. The outcome of this study is expected to assist the agrotechnology industry in early detection of paddy leaf disease in which an appropriate action could be taken accordingly.