Identification of Healthy and Diseased Paddy Leaves using Texture Features with ECOC Classifier

ABSTRACT Healthy and diseased leaves identification in agriculture field is the most challenging task. Earlier the experts use to visit the agricultural fields to observe or identify the disease which occurs on the plant surface which was time-consuming and expensive to the agriculturalists. In this proposed work, identification of diseased and healthy leaves is carried out using local binary pattern features. Error Correction Output Code (ECOC) classifier is used to classify the data, and obtained success rate of 87.60%. Smoothening of images is carried out with interpolation of scalar values for the RGB components.