‘Super Fuzzy’ Feature Model for Defects Detection of Fabric Materials
Aiming at many problems of existing feature model of fabric defects detection, such as large calculation, not exact expression and not extensive defects categories, a new ‘super fuzzy’ feature is proposed in this paper. This new method can solve above problems. Firstly, a general model of ‘super fuzzy’ is given. Some cent-characteristics denoted different kinds of defects are combined according to weight factor, and each cent-characteristic is guaranteed in a same quantity grade by modified constant vector. Then, a concrete ‘super fuzzy’ feature model combined by four cent-characteristics for fabric materials is obtained, and this model is optimized. Finally, after some programs experiments, concrete effect of fabric materials defects denoted by ‘super fuzzy’ feature is validated. Results show that ‘super fuzzy’ algorithm is effective for fabrics defects detection with different texture considering fabric characteristics, and defects feature of many fabrics is expressed quickly and exactly with no pre-supervised learning.
Z. Liu "‘Super Fuzzy’ Feature Model for Defects Detection of Fabric Materials", Advanced Materials Research, Vols. 108-111, pp. 872-877, 2010