Optimizing Woven Fabric Defect Detection using Image Processing and Fuzzy Logic Method at PT. Buana Intan Gemilang

Abstract

The development of textile industry which 3rd position in the largest export values in Indonesia prove that the quality of textile must be one of factors that should be considered for all of textile companies. Buana Intan Gemilang is one of the companies that produce woven fabric. This company s produce curtain woven fabric. Quality is the most important factor to reach high level customer satisfaction. To get the best quality of product needs to consider their quality control. According to inspection process in Buana Intan Gemilang, manual inspection for woven fabric defect detection need 19.87 second for average inspection time. Therefore, unbalance of production volume with inspection process cause the bottle neck in inspection process. In this research, proposed designing automated fabric inspection using image processing and Fuzzy Logic Model. This processed uses GLCM as feature extraction with three parameters are cluster shade, cluster prominence, and number of object. The proposed fabric inspection using Fuzzy Logic implemented with MATLAB provides better result in identifying fabric defect and optimizing process time. This research using 120 training data, 80 offline test data, and 80 real time test data. Identification automation defect of woven fabric test data can produce accuracy 97, 5% and averaging process time 1.15 second. This result is better than manually inspection process that took 19.87 second for scanning defect of woven fabric.