摘要
织物性能的分类、分级信息是纺织品质量评估的直观表述。在织物客观评估的基础上,提出了基于减法聚类与模糊C 均值(FCM)聚类的集成方法用于纺织品质量评估分析。该方法以减法聚类算法得出的样本的最佳分类数为基础,用FCM聚类算法得到具体的分类结果。将聚类中心的特征值之和定义为分级指数,进一步用于解决织物质量的分级问题。通过对法国鲁贝高等纺织工程学院自织的43块棉针织物的分析,证明了以上方法在处理纺织品质量分类、分级问题中的有效性。
The performance of textile products is a straightforward expression of their quality evaluation. In this paper, based on objective evaluation of textile products. An integrated method was proposed by combining subtractive clustering and fuzzy C-means (FCM) clustering algorithms. In the method, after obtaining the best number of classes for the textile samples by using the subtractive clustering algorithm, the clustering results can be further classified by using FCM algorithm. By defining the grading factor, which is the sum of eigenvalues of the clustering centers, the problem of objective grading could be solved. The effectiveness of the method in textile clustering and grading is proved by using 43 cotton samples which are self-knitted by Ecole Nationale Supérieure des Arts et Industries Textiles, France.
出处
《东华大学学报（自然科学版）》
CAS
CSCD
北大核心
2005年第1期54-58,共5页
Journal of Donghua University(Natural Science)
基金
国家自然科学基金(60474037
60004006)
新世纪优秀人才支持计划
教育部高等学校博士点专项基金(20030255009)
