混纺麻纱特征的实用图像表征技术——I.截面图像生成与特征指标

混纺织物的纤维组成和混纺比早在成纱时就已确定,其预先判定极为必要。本文探讨图像技术测量麻/涤混纺纱混合比的快速算法及其实用指标。该表征从麻/涤纱的有效切片、纱线截面摄像、图像小波分析去噪开始,到采用形态滤波法,解决图像局部灰度不均匀;使用分水岭法,分开粘连纤维和减少无效分割,实现图像的采集和预处理。实验结果证明,切片采样和小波除噪可达清晰采像要求;图像预处理可有效提高图像质量和分析精度。由此对纤维截面几何特征分析,得出主要特征参数为纤维当量截面积、异形系数和中腔纹。并依此参数及其组合完成对苎麻和涤纶纤维的自动识别,准确率达99.5%.

Applied Characterization for Ramie Blended Yarns Based on Image Analysis Part I.The Image Formation and Typical Parameters of Fibre Cross-sections

It is vitally important to find the blend ratio and component of fabrics because those have been fixed before yarn processing.The quick algorithms and the possibility for standard measurement of the blend ratio of ramie/polyester blended yarns are investigated in terms of image process technology in these series papers(Part I and II).The new practical image method includes yarn sectioning,digital photographing and noise deducting by wavelet-analyzing method to take effective images.Meanwhile,the gray filtration is adopted to uniform the whole image,and the watershed segmentation is used to separate the fibers clinging together and to reduce the ineffectual segmentation.The characteristic parameters,i.e.cross-sectional area,profiled coefficient and lumen line of single fibres,have been obtained according to fibre geometrical analysis.Through the individual and combined parameters,the auto-recognition to these two component fi-bres can be realized and its accuracy achieves to99.5%.