Ega Noviastuti
Program Studi Matematika, Universitas Negeri Yogyakarta

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

Model Identifikasi Singkong Berdasarkan Warna untuk Tepung Mocaf Berbasis Citra Digital Sri Andayani; Ega Noviastuti
Pythagoras: Jurnal Matematika dan Pendidikan Matematika Vol 17, No 1: June 2022
Publisher : Department of Mathematics Education, Faculty of Mathematics and Natural Sciences, UNY

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21831/pythagoras.v17i1.34994

Abstract

Penelitian ini bertujuan menghasilkan model untuk mengidentifikasi mutu singkong berdasarkan warna sebagai bahan pembuatan tepung mocaf dengan berbasis citra digital.  Metode yang digunakan meliputi beberapa tahap pengolahan citra digital antara lain preprocessing dan ekstraksi ciri. Preprocessing meliputi cropping, resizing, dan grayscaling, sedangkan ekstraksi ciri meliputi segmentasi threshold dan binerisasi. Data penelitian menggunakan 118 citra singkong yang dibagi menjadi 72 citra data latih dan 46 data uji. Hasil penelitian berupa model identifikasi yang mendasarkan pada dua hal berikut: a) menggunakan ekstraksi ciri yang meliputi segmentasi threshold dengan nilai ambang 170 dan binerisasi dengan nilai ambang 75; dan b) penentuan mutu singkong dilakukan berdasarkan perbandingan luas piksel putih hasil segmentasi threshold dengan luas piksel putih hasil binerisasi. Singkong dikatakan bermutu baik jika citranya yang memiliki persentase luas piksel warna putih lebih besar atau sama dengan 65%. Model yang dihasilkan memberikan akurasi sebesar 94% terhadap 72 data latih dan sebesar 95% terhadap 46 data uji. Cassava Identification Model Based on Color for Mocaf Flour Using Digital ImageAbstractThis study aims to produce a model to identify the quality of cassava-based on color as an ingredient for making mocaf flour based on digital images. The procedure includes preprocessing and feature extraction among other steps of digital image processing. Preprocessing includes cropping, resizing, and grayscaling, while feature extraction includes threshold segmentation and binaryization. The data are 188 cassava images consisting of 72 training data images and 46 test data. The result of the study is an identification model based on the following two things: a) utilizing feature extraction that uses threshold segmentation with a threshold value of 170 and binaryization with a threshold value of 75; and b) determining of the quality of cassava is carried out based on the ratio of the area of white pixels produced by threshold segmentation to the area of white pixels produced by binaryization. If 65% or more of the pixels in the image are white, the cassava has a good quality. The resulting model provides an accuracy of 94% for 72 training data and 95% for 46 test data.