Automatic Detection of Structure Parameters of Yarn-dyed Fabric

The structure parameters of woven fabric, such as density and woven pattern, are the most important information during the production. In the traditional manual method, these parameters are detected by the inspectors with a tex-tile magnifying glass. It is time-consuming and labor-inten-sive, and can not avoid the careless mistakes committed by the inspectors, which makes sample analysis become the bottleneck in the textile production process. With the development of computer technology, lots of methods based on image analysis have been proposed since 1980s last century to recognize the fabric density [1– 10] and pattern [1, 6, 11–21]. The methods can be divided into two main categories. (1) Frequency domain methods [1–4, 11–15]. The peaks in the Fourier transform results of woven fabric image represent the periodic structur