利用模糊C-均值聚类分析法实现织物组织结构自动识别

Automatic recognition of fabric weave patterns by fuzzy C-means clustering method

文章提出了一种新的有效的织物组织结构识别算法。用彩色扫描仪输入紧密织物组织的灰度图像并将其转换为数字文件,然后通过灰度图像形态学处理获得增强图像。基于纱线间隙和经纬纱交叉区的存在,通过一阶和二阶的统计量可获取4种区域结构特征。利用模糊C-均值聚类分析法得出识别经纬浮点的非监督的判别准则。实验材料包括平纹、斜纹和缎纹织物,实验结果表明这3种基础组织结构模式可以得到有效识别。

A new recognition algorithm is proposed for fabric weave pattern recognition. The gray-level image of solid woven fabrics is captured by a color scanner and converted into digital files, then enhanced images are obtained by a gray-leval morphological operation. Based on the interstices of yarns, warp and weft crossed areas are located, and four textures of these areas are obtained by first-order and second-order statistics. Unsupervised decision rules for recognizing warp and weft floats are developed using a fuzzy C-means clustering method. The experimental materials include plain, twill, and stain woven fabrics. Experimental results demonstrate that three basic weave patterns can be clearly identified.