Classification of Rice Leaf Spot Disease using Local Binary Patterns

Abstract

The fundamental objective of this work is to develop an image processing framework that can perceive a proper methodology for ContentBasedImageRetrieval(CBIR) in Leaf Inadequacy. The salient point selection concept is utilized by selecting the Salient points from the edgy image and the concept of inter-plane relationship method is imposed, LocalBinaryPatterns (LBPs) are computed with respect to the center pixel of the salient point. The research work consists primarily of three sections, namely representation of the leaf image, extraction of features and classifying. During the extraction process of the application the most important and special features of the image are retrieved. The image is contrasted with the data base images in the classification phase. The surface of the plant leaf is divided into smaller regions using which the LBP is obtained and the combination of them produces a single feature vector. An accurate model is constructed by this feature vector which is used to measure differences between flawed and healthy plant images.