Open AccessParijoto Fruits Classification using K-Nearest Neighbor Based on Gray Level Co-Occurrence Matrix Texture Extraction
Author(s) -
Ibnu Utomo Wahyu Mulyono,
T C Lukita,
Christy Atika Sari,
De Rosal Ignatius Moses Setiadi,
Eko Hari Rachmawanto,
Ajib Susanto,
M. Dalvin Marno Putra,
Dewi Agustini Santoso
Publication year - 2020
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1501/1/012017
Subject(s) - gray level , artificial intelligence , pattern recognition (psychology) , normalization (sociology) , histogram , co occurrence matrix , pixel , computer science , k nearest neighbors algorithm , texture (cosmology) , mathematics , computer vision , image processing , image (mathematics) , image texture , sociology , anthropology
Parijoto fruit is a typical fruit that grows around Mount Muria, Kudus Regency and Mount Merapi in Yogyakarta, Indonesia. This fruit has many health benefits, especially for pregnant women. This fruit production is not much because of its limited growth around Mount Muria. So parijoto fruit is made into powder drink products and syrup, so that it can be consumed in a longer period of time and not only during the harvest. To get a good processed product requires good quality ingredients. Parijoto fruit needs to be sorted and classified. Current technology allows classification to be done by digital image processing. The Gray Level Cooccurrence Matrix (GLCM) method is proposed to extract the texture features from the parijoto fruit and then classify them using the K-Nearest Neighbor (KNN) method. GLCM can describe a spatial linear relationship of the frequency at which gray values are determined by other gray values in one area of investigation. It can simply use the statistical approach of appearance or histogram of the image matrix. In this way, information will be easily relative position of neighboring pixels that are suitable for the classification process using KNN. KKN was chosen because this method was proven to be used for relatively few datasets, but a normalization process was needed to increase accuracy. Based on the results of the implementation of the GLCM and KNN methods for parijoto fruit classification the classification accuracy was 80%.