Measurement of the Severity of Rice Blast Disease by using Image Processing

Abstract

Rice blast disease is one of the main factors influencing crop yield. This researchis conducted to automatically measure the severity of rice blast disease that appears on the rice leaf by using image processing. The symptoms of this disease can represent the severity of the disease, which is used to select a suitable method for disease control. The severity of the disease is divided into four levels, such as no lesions, slight, moderate, and massive lesions. Different severity levels require the different techniques for controlling the disease, for example, the farmer should use the chemicals when the disease severity is at least at the moderate level.Rice leafs images were collected andput into image processing system for identifying the severity of rice blast disease by using K-Nearest Neighbor algorithm. This method can accurately identify around74.17%. Healthy leaf identification gives the best performance while the little lesions leaveshas appearance similar to healthy leaves. This work will be useful especially for farmers to estimate the disease severity and to select appropriate approaches for real-time disease management.