Klasifikasi Jenis Jambu Air Berdasarkan Bentuk Daun dengan Menggunakan Metode K-NN (K-Nearest Neighbor) Berbasis Android

Abstract

The classification of guava based on the shape of the leaves has an important role in an introduction to plant species due to plant diversity. One of them is a leaf plant of the water guava type. The method of Feature Extraction or known as feature extraction aims to classify images based on their similarity or shape similarity. In a study this time a classification system of types of guava leaves based on leaf shape will be created using the K-Nearest Neighbor (K-NN) method. In one study, there were 3 (three) types of guava leaves that would be classified, namely red guava leaves, red button guava leaves, deli honey guava leaves. And in this research the data used were 300 images, each leaf consisting of 100 leaf images for training data, while for data testing it used 75 image data, each leaf consisting of 25 leaf images. After conducting a test, it can be concluded that using the K-NN method is quite good at classifying 3 types of guava leaves and obtaining an accuracy value of 92%.