Layout Classification of Red Onion Disease on Onion Leaf Image Using Artificial Neural Network

Abstract

Abstract Shallots are one of the main ingredients that are often found throughout Indonesia. The optimal yield of onion plants is influenced by several factors, one of which is disease. The purpose of this study is to classify the image of leeks to identify onion diseases using Artificial Neural Network (ANN). Disease classification on leeks is leaf rot (anthracnose) and purple spots. The steps taken in the classification of leek disease are data input, pre-processing, feature extraction, machine learning. Several steps are carried out on the leek image, among others, improving image contrast, conversion of sRGB to LAB, segmentation using clustering, grayscale images, and binary images. After pre-processing, the next step is extracting features based on colour features and texture features. The colour features consist of Standard Deviation, Kurtosis, Mean, and Skewness. While the texture features consist of Contrast, Correlation, Energy, Entropy, Variance, and IDM. The end result of this system will show whether the leeks are included in the class of leaf rot (anthracnose) or purple spots or healthy leaves.