基于计算机成像的棉花伪异性纤维识别检测方法研究  

Research on Identification and Detection Method of Cotton Pseudo-foreign Fiber Based on Computer Imaging

作　　者：纪晓杰 王波 JI Xiao-jie;WANG Bo(Qingdao Institute of Textile Fibre Supervision and Inspection,Qingdao 266061 China)

机构地区：[1]青岛市纤维纺织品检验研究院,山东青岛266061

出　　处：《自动化技术与应用》2024年第11期68-70,87,共4页Techniques of Automation and Applications

摘　　要：针对现有棉花伪异性纤维识别检测方法中存在的分割能力较弱,导致棉花伪异性纤维图像灰度值与阈值相差较大。提出棉花伪异性纤维识别检测方法。利用图像分析棉花伪异性纤维,引入小波变换方法降低图像噪声值;采用最大类间方差算法完成模糊伪异性纤维图像的分割。采用自适应方法完成伪异性纤维特征分类,将相同染色体的纤维特征因子划分在同一区域,提取棉花伪异性纤维特征;根据计算图像内棉花伪异性纤维因子出现概率,实现识别检测。实验结果表明:所提方法可有效提高棉花伪异性纤维图像的分割能力,缩小灰度值与阈值之间差距。In view of the weak segmentation ability of the existing cotton pseudo foreign fiber identification and detection methods,resulting in a large difference between the gray value and the threshold value of the cotton pseudo foreign fiber image,a cotton pseudo foreign fiber identification and detection method is proposed.Using the image analysis of cotton pseudo foreign fiber,the wavelet transform method is introduced to reduce the image noise value,and the maximum inter class variance algorithm is used to complete the segmentation of fuzzy pseudo foreign fiber image.The self-adaptive method is used to complete the pseudo foreign fiber feature classification.The fiber feature factors of the same chromosome are divided into the same region,and the cotton pseudo foreign fiber features are extracted.According to the probability of occurrence of the false foreign fiber factor in the image,the recognition and detection is realized.The experimental results show that the proposed method can effectively improve the segmentation ability of cotton pseudo foreign fiber image,and narrow the gap between gray value and threshold value.

关 键 词：计算机成像 棉花伪异性 纤维识别 图像分割 纤维检测 

分 类 号：TP391.413[自动化与计算机技术—计算机应用技术]