应用显著性算法的纱线条干均匀度检测

发布时间：2018-03-30 15:30

本文选题：纱线条干均匀度　切入点：显著性分析　出处：《纺织学报》2017年06期

【摘要】：针对运用图像方法进行纱线条干均匀度检测时,背景黑板、纱线毛羽以及图像噪声等对检测结果影响较大的问题,借鉴人的视觉感知机制,提出一种应用显著性算法检测纱线条干均匀度的方法。对采集到的纱线图像提取颜色和亮度特征,进行显著性分析,突出纱线条干区域,然后利用迭代阈值分割算法和区域滤波,得到准确清晰的纱线条干二值图像,基于此进行直径计算、均匀度分析和纱线疵点判定。通过边缘准确性评价可知,采用所提方法分割得到的纱线条干二值图像有着较高的分割精度。通过与Uster Classimat 5的均匀度检测结果进行比较,证明这种方法可得到准确的结果,与Uster Classimat 5的测量结果有着较好的一致性。
[Abstract]:In order to solve the problem that background blackboard yarn hairiness and image noise have great influence on yarn evenness detection using image method for yarn evenness detection this paper draws lessons from human visual perception mechanism. A method of detecting yarn evenness by using salience algorithm is proposed. The color and brightness characteristics of the collected yarn images are extracted, and the salience analysis is carried out to highlight the yarn dry area. Then using iterative threshold segmentation algorithm and region filtering to obtain accurate and clear yarn yarn dry binary image, based on which the diameter calculation, evenness analysis and yarn defect determination are carried out. The yarn dry binary image obtained by the proposed method has higher segmentation accuracy. By comparing with the results of Uster Classimat 5, it is proved that this method can get accurate results. The results are in good agreement with that of Uster Classimat 5.
【作者单位】：西安工程大学电子信息学院;
【基金】：国家自然科学基金项目(61301276) 陕西省工业科技攻关项目(2015GY034)
【分类号】：TP391.41;TS101.922

【相似文献】