Design Of Experiments Application Using Taguchi Approach To Identify Woven Fabrics Defects By Image Processing At Cv. Maemunah Majalaya

Abstract

Grey fabric is one of the woven fabrics types that become the core product in CV. MAEMUNAH MAJALAYA. The inspection activity in this company is still conducted manually by using human vision. Therefore, that existing condition gives a very crucial implication in production flow. The limited numbers of inspection station cannot afford the production volume that increases in every week. Each station is only capable to inspect 30 meters of fabrics in one hour, while there are more than 20.000 meters of fabrics demand in every week. Therefore, automated system through image processing system for conducting the inspection activity can be implemented to optimize the productivity.

In order to support the image processing system, this study will conduct design of experiment (DOE) using Taguchi approach to get the required data or information to build the image processing application. The Taguchi approach is utilized different combinations of factors and levels. The chosen factors are 463 lux light intensity, 20 cm camera distance, 8 MP camera resolution, 0.2 threshold, and 47 grayscale with low, medium, and high as the levels for each factors. Therefore, through Taguchi Orthogonal Array there will be 27 of experiments combinations will be discussed in this study

Keywords: Design of Experiment, Taguchi Approach, Image Processing System, Woven Fabrics, Fabrics Inspection, Orthogonal Arrays.