Identification of rice leaf disease based on rice leaf image features using the k-Nearest Neighbour (k-NN) technique

Abstract

Increasing productivity of rice plants is crucial to offset the rate of population growth because rice for most of the world's population is the primary energy source. The phenomenon of degradation of fertility and disease in rice plants poses a severe challenge, prevention and control measures are needed. The health of rice is the main factor that influences productivity. Diseases of rice leaves include various fungal pathogenic diseases such as rice blast, brown spots, and leaf blight. It is difficult to identify the type of rice leaf disease. This study discusses a digital image processing model for classifying rice leaf disease use leaf image features. Experiments conducted in this study used three types of rice leaf diseases, namely rice blast, brown spots, and leaf blight. The k-Nearest Neighbour algorithm was used as the primary technique to classify the image based on its features such as features of shapes, patterns, and feature colors. The results of the experiment showed that the average accuracy performance was 77% for the precision and recall was 74%.