图像处理技术测定纱线混纺比——特征值选择的系统聚类和模糊C均值聚类

Research for the Measurement of Polyester/Cotton with Image Analysis——Feature Selection and Fiber Clustering Analysis

基于已有的特征指标,对涤和棉纤维的各特征值进行选择和分类。利用系统聚类和模糊C均值聚类方法,对特征指标进行了比较,最终选用异形度指标对纤维进行分类,实现了纱线混纺比的计算。研究表明,对于涤棉混纺纱线,异形度指标可以很好地反映纤维间的差异。模糊C均值聚类方法较系统聚类方法,能有效地将纤维区别开来。

Based on the characteristic parameters, feature selection and fiber clustering analysis are studied for the yarn of polyester/cotton, comparison among the characteristic parameters is discussed and finally the characteristic of fiber shaped degree is adopted to classify polyester/cotton. Results indicate that fiber shaped degree can be used to distinguish the blend yarn of polyester/cotton accurately. Compared with Hierarchical clustering analysis method, Fuzzy C-means cluster algorithm represents a significant advance and can be applied to fiber recognition more effectively.