Premium
Analysis and Simulation of Nonwoven Textures
Author(s) -
Scholz Michael,
Claus Bernhard
Publication year - 1999
Publication title -
zamm ‐ journal of applied mathematics and mechanics / zeitschrift für angewandte mathematik und mechanik
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.449
H-Index - 51
eISSN - 1521-4001
pISSN - 0044-2267
DOI - 10.1002/zamm.19990791362
Subject(s) - homogeneity (statistics) , artificial intelligence , markov random field , computer science , computer vision , pattern recognition (psychology) , image (mathematics) , image segmentation , machine learning
Nonwoven fabrics are widely used in a variety of different fields. They consist of bonded fibers which are more or less randomly distributed (in particular they are not woven), and in all applications the homogeneity of the fabric is an important feature which is critical to the quality of the fabric. We evaluate the homogeneity (cloudiness) by decomposing an image of a nonwoven into different scales, evaluating the homogeneity on each scale and combining these results into a single number which represents a measure of the overall „cloudiness” of the fabric in the image. Simulation of images of nonwovens with a predefined degree of homogeneity (on each of the considered scales) is of importance because these images can be used to calibrate the image acquisition system. In particular we use Markov Random fields to generate the characteristic texture for each scale after first identifying the respective model parameters. We then combine these different images to yield an „artificial” image of a nonwoven.