基于形态学处理的提花针织物纹样相似性识别  

Pattern similarity recognition of jacquard knitted fabrics based on morphological

作　　者：杨婷 胡佳琳 唐泽华 李欣欣 YANG Ting;HU Jialin;TANG Zehua;LI Xinxin(College of Textiles,Donghua University,Shanghai 201620,China)

机构地区：[1]东华大学纺织学院,上海201620

出　　处：《毛纺科技》2023年第11期85-90,共6页Wool Textile Journal

基　　金：国家自然科学基金项目(61902357);中央高校基本科研业务费专项资金资助项目(2232020D-12)。

摘　　要：为了提高针织工厂提花纹样相似性识别效率及准确率,避免传统人工识别的主观差异,提出基于形态学边缘检测的提花针织物纹样相似性识别算法。首先,利用工艺软件设计针织提花纹样并构建提花数据库,通过MatLab对纹样图像进行预处理,之后经腐蚀、膨胀等操作提取提花结构轮廓特征图像,基于相似性度量公式求解数据库中各纹样与目标纹样的相似性,筛选出相似率达到设定阈值之上的提花纹样。结果显示:在样本容量为100的纹样数据库中,目标样本筛选速度不超过1.21 s,且不易受主观因素的影响;基于图像处理的提花针织物纹样相似性识别法具有高效率和可靠性,可为生产企业快速筛选目标纹样提供技术支持。In order to improve the efficiency and accuracy of jacquard pattern similarity recognition in knitting factories and avoid the subjective differences of traditional manual identification,a jacquard knitted fabric pattern similarity recognition algorithm based on morphological edge detection was proposed in this paper.Firstly,the knitted jacquard database was designed and constructed by process software,the pattern image was prepossessed by MatLab,and then the outline feature image of the jacquard structure was extracted by corrosion and expansion,after then,the similarity between each pattern in the database and the target pattern was retrieve out by the similarity measurement formula.Finally,the jacquard pattern of the similarity rate above the set threshold was screened.The results show that in the pattern database with sample size of 100,the target sample screening speed is less than 1.21 s,and it is not easily affected by subjective factors.Image processing based jacquard knitted fabric pattern similarity recognition method has high efficiency and reliability,and can provide technical support for manufacturers to quickly select target patterns.

关 键 词：针织 提花结构 形态学 边缘检测 相似性识别 

分 类 号：TS18[轻工技术与工程—纺织材料与纺织品设计]