利用图像处理和分析技术测定羊毛/羊绒混纺比

叙述了利用图像处理技术测定羊毛/羊绒混纺纱混纺比的新途径.在预处理阶段,通过数学形态学操作对羊毛/羊绒混纺纱截面图像进行去噪和平滑.在人眼视觉机理的基础上,提出光斑扩散方法,用来探测图像中各个纤维截面并得到其轮廓线.在特征抽取阶段,构建了面积系数作为区分羊毛和羊绒纤维的特征指标.对抽取到的特征数据进行聚类和分类,累计每一类纤维的面积系数之和.最后,考虑到2种纤维的密度,计算出羊毛/羊绒混纺比.通过对同一羊毛/羊绒混纺纱的不同部分采集一系列截面图像进行测试,得到了可靠的结果.

Determination of wool/cashmere blending ratio by image processing and analysis

This paper introduces a new approach based on image processing techniques to determine blending ratio of the wool/cashmere blended yarn.In phase of preprocessing,the cross-section images of wool/cashmere blending yarn are denoised and smoothed by means of mathematical morphology operations.Based on mechanism of human sight,the facula diffusion method is advanced to detect individual cross-sections in the image and takes out the contours of them. In phase of feature extraction,the dimension coefficient is pu...