Design on inspection system automation of curtain woven fabrics by image processing and the taguchi approach method at PT. Buana Intan Gemilang

摘要

The textile industry is one of the industries that provide high export value by occupying the third position in Indonesia. The process of inspection on traditional textile enterprises by relying on human vision that takes an average scanning time of 19.87 seconds. Each roll of cloth should be inspected twice to avoid missed defects. This inspection process causes the build-up at the inspection station. This study proposes the automation of inspection systems using image processing and the Taguchi Approach Method to test the combination of factors that affecting the inspection process. The input for image processing and Taguchi Approach comes from the existing state and previous research as reference and trial and error process. The result of the Taguchi Approach is error rate and will be processed using S/N ratio and ANOVA, which will be determined which factor that has the most affecting the inspection process and determined the best combination (robust) for inspection process. The result showed combination of light intensity 13001x, camera distance 16 cm, camera resolution 1280x960 pixels, grayscale value 10, threshold value 0.3, and edge detection operator sobel that has small error in inspection process.