Abstract:

Fuzz on the fabric surface is very important for appearance quality, since it can lead to pilling and an unpleasant handle and appearance. Thus, it is important to predict the fuzz value on the fabric surface before producing the fabric. The amount of fuzz on the fabric surface, as has been determined by image processing techniques, can be predicted by two different methods, the Artificial Neural Network (ANN) and Regression Analysis. During knitting, the yarn used is abraded to an uncertain degree, depending on the knitting conditions and yarn properties, and this situation can lead to poor prediction results. However, it has been demonstrated that the prediction results obtained by Artificial Neural Networks look more promising than that of Regression Analyses.