摘要
针对基于常规特征指标分类时导致的化学纤维异型喷丝孔细小污垢误判问题,提出一种基于非线性孔轮廓曲线的模板匹配方法。从孔边缘轮廓曲线分布特征着手,构建包含控制点的参数化标准检测模板,再通过非刚性配准,将模板轮廓配准至待测孔边缘轮廓,构建异型孔归一化的污垢检测曲线,分割、定位污垢并构建污垢量化指标。采集经专家分类评估且包含多种污垢分布的扁平孔、三叶孔及十字孔等3种常规异型孔小型数据库,3个数据库的试验结果显示,基于非线性孔轮廓曲线模板匹配方法的污垢识别正确率达95%以上,显著高于传统方法。结果表明,该方法能够满足化学纤维常规异型喷丝孔的污垢检测要求,在喷丝板全自动机器视觉质量评估领域具有较好的应用前景。
To solve the problem of misjudgment of small dirt in chemical fiber special-shaped spinneret holes caused by classification based on conventional characteristic indexes,a template matching method based on nonlinear hole contour curve was proposed.According to the distribution characteristics of hole edge contour curve,a parameterized standard detection template containing control points was generated.Through non rigid registration,the template contour was registered to the profile to be measured,and a normalized dirt detection curve of special-shaped holes was calculated,and dirts were segmented and located,and the quantitative indicators of dirts were constructed.Flat-shaped,Y-shaped and cross-shaped types of small image database with various degrees of dirt distribution and the assessments of expert evaluation were collected.The experimental results of three databases show that the accuracy of the template matching method based on nonlinear hole contour curve is more than 95%,which is significantly higher than that of the traditional method.The proposed method can meet the requirements of dirt inspection for chemical fiber special-shaped spinneret holes,and has a good application prospect in the field of automatic machine vision quality assessment of spinnerets.
作者
徐律
XU Lü(School of Textile and Fashion,Shanghai University of Engineering Science,Shanghai 201620,China)
出处
《东华大学学报（自然科学版）》
CAS
北大核心
2023年第2期104-111,共8页
Journal of Donghua University(Natural Science)
关键词
异型喷丝孔
化学纤维
污垢检测
模板匹配
非刚性配准
special-shaped spinneret hole
chemical fiber
dirt inspection
template matching
nonrigid alignment
作者简介
通信作者:徐律,男,讲师,研究方向为机器视觉、模式识别等,E-mail:xulyu@sues.edu.cn。
