Open AccessOPTIMASI PROSES PENGUKURAN DIMENSI DAN DEFECT UBIN KERAMIK MENGGUNAKAN PENGOLAHAN CITRA DIGITAL DAN FULL FACTORIAL DESIGN
Author(s) -
Denny Sukma Eka Atmaja Muhammad Kusumawan Herliansyah
Publication year - 2015
Publication title -
jurnal teknosains: jurnal ilmiah sains dan teknologi/jurnal teknosains : jurnal ilmiah sains dan teknologi
Language(s) - English
Resource type - Journals
eISSN - 2443-1311
pISSN - 2089-6131
DOI - 10.22146/teknosains.7972
Subject(s) - tile , mathematics , ceramic , optics , geometry , materials science , physics , composite material
Theinspection process of surface quality of ceramic tile could be done by using image processing technique throughthe optimization by using Center for Ceramics’s parameteron Indonesian National Standard (SNI) ISO 10545.This research will analyze from light intensities (level 300lx, 600lx, and 900lx), and camera distances (50cm, 75cmand 100cm), with three times replication using full factorial design. This research uses Matlab 2009a softwareto identify area and defect on dry spots ceramic tile’s surface. The result obtained from this research is there weresignificant influencing factors: light intensity, and camera distance, as well as the interaction of these factorstowards the error rate percentage of measuring areaand defect on ceramic tile’s surface. The smallest error ratevalue from measuring tile’s surface and diameter of dry spots with light intensity of 300lx and camera distance of50cm had been obtained the error rate value for each measurement about 0.0675% and 2.30%, with combinationof grayscale value for the error rate measurements of tile’s surface and diameter of dry spots were 0.2989 x 0.1140x R + G+0.5870 x B. Based on the correlation coefficient value between light intensity, camera distance towardsthe error rate of measuring areaand defect on tile’s surface, each of them was obtained correlation coefficient valueof camera distance with error rate had 0.518 and 0.516, which meant a strong correlation. The positive correlationcoefficient value showed a unidirectional relationship of two variables, where the rise of one variable would causethe rise of another variable and the decline of one variable would cause the decline of another variable.