Rice diseases classification using feature selection and rule generation techniques
Development of an automation system for classifying diseases of the infected plants is a growing research
area in precision agriculture. The paper aims at classifying different types of rice diseases by extracting
features from the infected regions of the rice plant images. Fermi energy based segmentation method
has been proposed in the paper to isolate the infected region of the image from its background. Based
on the field experts’ opinions, symptoms of the diseases are characterized using features like colour,
shape and position of the infected portion and extracted by developing novel algorithms. To reduce complexity
of the classifier, important features are selected using rough set theory (RST) to minimize the loss
of information. Finally using selected features, a rule base classifier has been built that cover all the diseased
rice plant images and provides superior result compare to traditional classifiers