Implementation of CBIR Method for Identification of Corn Disease Using Extreme Learning Machine

Abstract

Various artificial neural network methods and digital image processing techniques have been applied in agriculture. Previous researchers have proposed methods for identification or detection of plant diseases, such as using K-NN, support vector machine (SVM), backpropagation and so on. The disease attack can reduce the amount of agricultural productivity and result in other material losses for farmers. Plant diseases can be observed from the physical and color changes of the affected parts such as roots, stems and leaves. However, with the development of computer vision technology, observations can be assisted by image processing methods and artificial neural networks, for automation in agriculture. This study proposes a method of applying digital image processing with artificial neural networks to identify corn plant diseases. The purpose of this research is as a comparative study of the proposed method with other conventional methods. The methods used are content based image retrieval (CBIR) and extreme learning machine (ELM). Content based image retrieval (CBIR) is the process of searching for an image in a database by comparing the features in the query image with the features that have been stored in the image database. Meanwhile, extreme learning machine (ELM) is an artificial neural network with one hidden layer or single hidden layer feedforward neural network (SLFNs). The feature extraction used consists of 3 features, namely color, shape, and texture feature extraction. The dataset consists of 3 types of images, namely images of healthy leaves, leaf spot disease, and rust disease. The ELM artificial neural network serves as a classifier that classifies the types of diseases based on the given feature extraction. The proposed system is able to identify the type of disease based on the pattern that appears on the leaf surface with the highest precision of 0.9922.