Özet:

In the textile industry, the detection of production errors on raw fabric is one of the most important and costly work done in the quality control phase. The fabric is checked before it comes to the confection phase and the wrong parts are taken by cutting into pieces from the fabric. This leads to a decline in production. If there are defects that have not been noticed during the check or that have not been separated into the fire with the collection of pieces, these clothes fall to the second quality. The quality checking after the ball cutting process of the defective piece of fabric or the second quality products sent to the customer without notice causes serious costs. The quality control process made in human sight, although it is quite expensive in terms of workforce and time, its success is not as high as desired. Therefore, in the quality control stage, work is done for the use of automatic and intelligent systems. In this study, image processing methods are recommended for the detection and classification of errors in denim tissues that are produced from 100% cotton thread and painted with indigo paint. In the study, the image processing algorithms were used to detect errors from the fabric images, with the extract of features, the number of data were obtained. The created character matrix provides a distinction between the erroneous and erroneous image studied in four different algorithms. The best performance algorithm is the J48 decision algorithm.

Anahtar Kelimeler: