【作者】 薛爱萍；

【导师】 刘长江；

【作者基本信息】 北京工业大学 ， 机械制造及其自动化， 2001， 硕士

【摘要】 纺织原料及其产品中纤维种类的鉴别是以各类纺织用纤维的形态特征及其理化性能为客观依据的，纤维细度和含量是重要的指标。传统上对羊绒及其混合纤维的检验与测试多利用投影显微镜法，这种方法人为因素干扰大，准确率不高，且劳动强度大，因此开发一套羊绒和羊毛混合纤维自动检测系统，对客观评定纤维的品质及数量具有重大意义。 本文论述了羊绒和羊毛混合纤维定量分析系统中的图象处理，特征提取和软件设计等问题。系统的分析过程为：待测量纤维经显微镜成象后，由CCD将光信号转换成电信号，再经图象卡将显微视场中的一块区域转化成一幅数字图象，然后用计算机对数字化纤维图象做分割处理（即二值化处理），找出待分析的纤维，通过快速纤维检测算法进行分析计算，测定各纤维组分的平均直径和根数等主要参数，从而计算出各纤维组分的含量。 本文针对纤维图象中的特点，提出了噪声抑制的快速中值滤波方法，这种方法对纤维边缘伪边具有良好平滑作用；在图象特征提取中，本文提出了基于方向信息的边缘检测算法，实验表明这种算法对纤维边缘定位精度高、检测出的边缘轮廓清晰，有利于下一步的测量计算；提取纤维图象特征时本文设计了测量直径的主轴法，这种方法对十到四十微米的纤维都能够准确测量。此外，本文还对步进电机驱动接口程序和软件界面提出了改进和完善措施。更多还原

【Abstract】 Shape and texture features of fiber play an very important role in distinguishing different textile materials. Among all these features, fiber diameter and content are two commonly used indicators. Conventional methods for fiber differentiation, which are often carried out by hand touch and eyeballing, have the disadvantages of low accuracy, bad robustness and inefficiency. Therefore, developing a set of intelligent automatic analyzer for fiber quality inspection is of great significance.In this dissertation, an intelligent analyzer based on image processing technology is developed for cashmere fiber quality inspection. The setup of the intelligent analyzer includes a personal computer, a image capturing card, an optical microscope, a CCD camera, a X-Y driving worktable, a printer and a set of software for image capturing and analyzing. Microscopic images are captured by the CCD camera and are saved in the hard disk in digital form. These images are then processed with the software, including the preprocessing of fiber image and feature extraction and classification. Finally, the diameter and content of cashmere fiber can be calculatedDiscussed in this dissertation are mainly the algorithm of median filter, edge extraction with orientation information and the methods of feature extraction.更多还原