基于纹理滤波和颜色聚类的提花织物纹样自动提取方法  

Automatic pattern extraction of jacquard fabric based on texture filtering and color clustering

作　　者：傅艺扬 刘妹琴[1,2] 樊臻 张森林[1] FU Yiyang;LIU Meiqin;FAN Zhen;ZHANG Senlin(College of Electrical Engineering,Zhejiang University,Hangzhou 310027,China;Lianshui Tiangong Fabric Manufacture Co.,Ltd.,Huai’an 223464,China)

机构地区：[1]浙江大学电气工程学院,杭州310027 [2]涟水天宫云锦织造有限公司,江苏淮安223464

出　　处：《丝绸》2019年第12期9-15,共7页Journal of Silk

基　　金：江苏省科技支撑计划项目(BE2014099)

摘　　要：为了提高提花织物纹样提取的准确性,消除织物组织结构对提取结果的干扰,文章提出了一种基于纹理消除滤波算法和密度峰聚类算法的纹样提取方法。通过统计图片区域梯度信息分离提花织物组织纹理区域和图案边缘结构,并利用非极大值抑制获得纹样边缘结构作为滚动引导滤波器的引导图,经过多次迭代计算实现织物纹理滤波。然后将图片从RGB颜色空间转换为CIELab颜色空间,利用密度峰聚类算法对织物色彩空间分割聚类,最终提取出织物纹样。实验结果表明,文章提出的方法快速准确地实现了提花织物纹样的自动分割与提取。This paper proposes a pattern extraction method based on texture elimination filtering algorithm and density peak clustering algorithm in order to improve the accuracy of jacquard fabric pattern extraction and eliminate the interference of woven fabrics structure on the extraction result. Firstly, the fabric weave texture region was separated from pattern edge structure by the statistics of regional gradient information of pictures. Secondly, the edge structure of patterns obtained by non-maximum suppression was used as the guidance image of rolling guidance filter to achieve fabric texture filtering by multiple iterative computations. Finally, the images were transported to CIELab color space from RGB color space, and density peak clustering algorithm was utilized to fast segment the fabric colors to gain the pattern extraction results. The experimental results show that the proposed method can fast and precisely segment and extract jacquard fabrics.

关 键 词：提花织物 滚动引导滤波器 密度峰聚类 图像分割 颜色提取 

分 类 号：TS941.2[轻工技术与工程—服装设计与工程] TS101.91